FIX: remove the docker socket from so-telegraf

so-telegraf mounted /var/run/docker.sock and joined the host docker group on every node type. The :ro flag blocks write() to the inode, not connect() plus HTTP over the socket, so any code execution inside the container could reach POST /containers/create with Privileged:true and become root on the host. No telegraf script used the socket; the only consumer was the native [[inputs.docker]] plugin, and group_add 920 existed solely to feed it. Container metrics now come from so-container-stats, a collector that runs on the host from cron and writes influx line protocol to a file telegraf already had mounted. This is the pattern so-status, so-raid-status and so-elasticagent-status already use, so the privileged docker access stays on the host side where root cron already ran it. The collector runs as somon, a service account in the docker group with no login shell and a locked password, rather than root. Docker group membership is still root-equivalent on the host, so this is defense in depth rather than a privilege boundary. The telegraf scripts were root:939 mode 770, letting socore rewrite them for code execution inside the container; they are now 750, which still allows the read and execute telegraf needs. The container also ran with no group, giving it gid 0, and now runs as 939:939. That alone would have broken lasthighstate.sh, which reached /opt/so/log/salt only via the root group and could not tell an unreadable file from a missing one, so it silently reported a 56 year highstate age. Only the lasthighstate file is bind mounted now, and the script tests readability instead of existence. Everything inputs.docker collected beyond the five fields the shipped dashboards query is available per stat under telegraf:container_stats, annotated for SOC so an operator can enable it without editing files. Defaults reproduce the previous output exactly. With every stat enabled the emitted field set matches what inputs.docker wrote, verified by running the plugin against the live socket and diffing: 53 fields, no type mismatches, no field present on one side only. Two deliberate differences: max_usage carries the real cgroup peak where the daemon reports 0 on cgroup v2, and host-network containers emit no docker_container_net row, matching inputs.docker. Docker label tags are not restored, since nothing queries them and they cost significant cardinality. so-status, the influxdb size cron, so-elasticagent-status, so-raid-status and so-common-status-check truncated their output in place while telegraf read it, so telegraf periodically saw an empty file and logged a parse error or emitted empty values. They now write aside and rename. Measured on a live manager, the old so-status cron left status.log empty for 215 of 10997 reads. Tested on a fresh install, a converted grid and a 3.0 upgrade.
Merge pull request #16260 from Security-Onion-Solutions/reyesj2/16254
2026-09-30 11:37:16 +02:00 · 2026-09-24 14:30:25 -04:00 · 2026-09-23 08:58:15 -05:00 · 2026-09-23 08:51:11 -05:00 · 2026-09-22 21:25:50 -05:00 · 2026-09-22 16:11:56 -05:00
58 changed files with 2085 additions and 138 deletions
@@ -6,6 +6,9 @@ on:
      - "salt/sensoroni/files/analyzers/**"
      - "salt/manager/tools/sbin/**"
      - "salt/_beacons/**"
+      - "salt/telegraf/tools/sbin_jinja/**"
+      - "salt/telegraf/defaults.yaml"
+      - "salt/telegraf/soc_telegraf.yaml"

 jobs:
  build:
@@ -34,3 +37,25 @@ jobs:
    - name: Test with pytest
      run: |
        PYTHONPATH=${{ matrix.python-code-path }} pytest ${{ matrix.python-code-path }} --cov=${{ matrix.python-code-path }} --doctest-modules --cov-report=term --cov-fail-under=100 --cov-config=pytest.ini
+
+  telegraf-collector:
+    # so-container-stats is a jinja template rather than an importable module, so it gets its
+    # own job: the test renders it the way salt does, then drives it with a faked docker engine
+    # and cgroup tree. No container runtime is needed.
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v3
+    - name: Set up Python
+      uses: actions/setup-python@v3
+      with:
+        python-version: "3.14"
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        python -m pip install flake8 pytest jinja2 pyyaml
+    - name: Lint with flake8
+      run: |
+        flake8 salt/telegraf/tools/sbin_jinja/so-container-stats_test.py --config=pytest.ini
+    - name: Test with pytest
+      run: |
+        pytest salt/telegraf/tools/sbin_jinja/so-container-stats_test.py -v
@@ -217,9 +217,11 @@ sostatus_log:
    - replace: False

 # Install sostatus check cron. This is used to populate Grid.
+# telegraf reads status.log on the same minute boundary this runs, so write aside and rename
+# rather than truncating the file it is reading
 so-status_check_cron:
  cron.present:
-    - name: '/usr/sbin/so-status -j > /opt/so/log/sostatus/status.log 2>&1'
+    - name: '/usr/sbin/so-status -j > /opt/so/log/sostatus/status.log.tmp 2>&1; mv -f /opt/so/log/sostatus/status.log.tmp /opt/so/log/sostatus/status.log'
    - identifier: so-status_check_cron
    - user: root
    - minute: '*/1'
@@ -9,6 +9,7 @@ import sys
 import subprocess
 import os
 import json
+import tempfile

 sys.path.append('/opt/saltstack/salt/lib/python3.10/site-packages/')
 import salt.config
@@ -17,6 +18,21 @@ import salt.loader
 __opts__ = salt.config.minion_config('/etc/salt/minion')
 __grains__ = salt.loader.grains(__opts__)

+def write_atomic(path, value):
+  # telegraf reads these files on its own schedule; replacing them by rename means it never
+  # reads a truncated file and reports an empty value as if it were real
+  directory = os.path.dirname(path)
+  handle, temp = tempfile.mkstemp(dir=directory)
+  try:
+    with os.fdopen(handle, 'w') as f:
+      f.write(str(value))
+    os.chmod(temp, 0o644)
+    os.replace(temp, path)
+  except Exception:
+    os.path.exists(temp) and os.unlink(temp)
+    raise
+
+
 def check_needs_restarted():
  osfam = __grains__['os_family']
  val = '0'
@@ -34,8 +50,7 @@ def check_needs_restarted():
  else:
    fail("Unsupported OS")

-  with open(outfile, 'w') as f:
-    f.write(val)
+  write_atomic(outfile, val)

 def check_for_fps():
    feat = 'fps'
@@ -56,8 +71,7 @@ def check_for_fps():
          # Unknown, so assume 0
          fps = 0

-    with open('/opt/so/log/sostatus/fps_enabled', 'w') as f:
-       f.write(str(fps))
+    write_atomic('/opt/so/log/sostatus/fps_enabled', fps)

 def check_for_lks():
    feat = 'Lks'
@@ -80,8 +94,7 @@ def check_for_lks():
                                lks = 1
        if lks:
            break
-    with open('/opt/so/log/sostatus/lks_enabled', 'w') as f:
-       f.write(str(lks))
+    write_atomic('/opt/so/log/sostatus/lks_enabled', lks)

 def fail(msg):
  print(msg, file=sys.stderr)
@@ -125,4 +125,6 @@ else
  RAIDSTATUS=1
 fi

-echo "nsmraid=$RAIDSTATUS" > /opt/so/log/raid/status.log
+# telegraf reads this file; write aside and rename so it never sees a half-written file
+echo "nsmraid=$RAIDSTATUS" > /opt/so/log/raid/status.log.tmp
+mv -f /opt/so/log/raid/status.log.tmp /opt/so/log/raid/status.log
@@ -1,6 +1,12 @@
 docker:
  range: '172.17.1.0/24'
  gateway: '172.17.1.1'
+  networks:
+    sobridge: {}
+    soauth:
+      range: '172.17.2.0/24'
+      gateway: '172.17.2.1'
+      manager_only: True
  ulimits:
    - name: nofile
      soft: 1048576
@@ -58,18 +64,18 @@ docker:
      ulimits: []
    'so-kratos':
      final_octet: 28
+      networks: ['soauth']
      port_bindings:
        - 0.0.0.0:4433:4433
-        - 0.0.0.0:4434:4434
      custom_bind_mounts: []
      extra_hosts: []
      extra_env: []
      ulimits: []
    'so-hydra':
      final_octet: 30
+      networks: ['soauth']
      port_bindings:
        - 0.0.0.0:4444:4444
-        - 0.0.0.0:4445:4445
      custom_bind_mounts: []
      extra_hosts: []
      extra_env: []
@@ -128,6 +134,7 @@ docker:
      ulimits: []
    'so-soc':
      final_octet: 34
+      networks: ['sobridge', 'soauth']
      port_bindings:
        - 0.0.0.0:9822:9822
      custom_bind_mounts: []
@@ -1,8 +1,26 @@
 {% import_yaml 'docker/defaults.yaml' as DOCKERDEFAULTS %}
 {% set DOCKERMERGED = salt['pillar.get']('docker', DOCKERDEFAULTS.docker, merge=True) %}
-{% set RANGESPLIT = DOCKERMERGED.range.split('.') %}
-{% set FIRSTTHREE = RANGESPLIT[0] ~ '.' ~ RANGESPLIT[1] ~ '.' ~ RANGESPLIT[2] ~ '.' %}
+
+{% if DOCKERMERGED.networks.sobridge is not mapping %}
+{%   do DOCKERMERGED.networks.update({'sobridge': {}}) %}
+{% endif %}
+{% do DOCKERMERGED.networks['sobridge'].update({'range': DOCKERMERGED.range, 'gateway': DOCKERMERGED.gateway}) %}
+
+{% for netname, net in DOCKERMERGED.networks.items() %}
+{%   set RANGESPLIT = net.range.split('.') %}
+{%   do net.update({'prefix': RANGESPLIT[0] ~ '.' ~ RANGESPLIT[1] ~ '.' ~ RANGESPLIT[2] ~ '.'}) %}
+{% endfor %}

 {% for container, vals in DOCKERMERGED.containers.items() %}
-{%   do DOCKERMERGED.containers[container].update({'ip': FIRSTTHREE ~ DOCKERMERGED.containers[container].final_octet}) %}
+{%   set CONTAINER_NETS = vals.get('networks', ['sobridge']) %}
+{%   set IPS = {} %}
+{%   for netname in CONTAINER_NETS %}
+{%     do IPS.update({netname: DOCKERMERGED.networks[netname].prefix ~ vals.final_octet}) %}
+{%   endfor %}
+{%   do DOCKERMERGED.containers[container].update({
+         'networks': CONTAINER_NETS,
+         'ips': IPS,
+         'network': CONTAINER_NETS[0],
+         'ip': IPS[CONTAINER_NETS[0]]
+       }) %}
 {% endfor %}
@@ -71,15 +71,19 @@ dockerreserveports:
    - source: salt://common/files/99-reserved-ports.conf
    - name: /etc/sysctl.d/99-reserved-ports.conf

-sos_docker_net:
+{% for NETNAME, NETWORK in DOCKERMERGED.networks.items() %}
+{% if not NETWORK.get('manager_only') or GLOBALS.get('is_manager', False) %}
+sos_docker_net_{{ NETNAME }}:
  docker_network.present:
-    - name: sobridge
-    - subnet: {{ DOCKERMERGED.range }}
-    - gateway: {{ DOCKERMERGED.gateway }}
+    - name: {{ NETNAME }}
+    - subnet: {{ NETWORK.range }}
+    - gateway: {{ NETWORK.gateway }}
    - options:
-        com.docker.network.bridge.name: 'sobridge'
+        com.docker.network.bridge.name: '{{ NETNAME }}'
        com.docker.network.driver.mtu: '1500'
        com.docker.network.bridge.enable_ip_masquerade: 'true'
        com.docker.network.bridge.enable_icc: 'true'
        com.docker.network.bridge.host_binding_ipv4: '0.0.0.0'
-    - unless: ip l | grep sobridge
+    - unless: ip l | grep {{ NETNAME }}
+{% endif %}
+{% endfor %}
@@ -7,6 +7,40 @@ docker:
    description: Default docker IP range for containers.
    helpLink: docker
    advanced: True
+  networks:
+    sobridge:
+      description: |
+        The default docker network, carrying most containers. Its range and gateway are taken
+        from the docker.range and docker.gateway settings above rather than set here.
+      helpLink: docker
+      readonly: True
+      advanced: True
+      global: True
+    soauth:
+      range:
+        description: |
+          IP range for the soauth docker network, an isolated network for the authentication
+          services, so that the Kratos and Hydra admin APIs are only reachable from the
+          containers placed on it.
+        helpLink: docker
+        readonly: True
+        advanced: True
+        global: True
+      gateway:
+        description: Gateway for the soauth docker network.
+        helpLink: docker
+        readonly: True
+        advanced: True
+        global: True
+      manager_only:
+        description: |
+          Limits the soauth network to grid members running the authentication containers,
+          instead of creating it on every node.
+        helpLink: docker
+        readonly: True
+        advanced: True
+        global: True
+        forcedType: bool
  ulimits:
    description: |
      Default ulimit settings applied to all containers via the Docker daemon. Each entry specifies a resource name (e.g. nofile, memlock, core, nproc) with soft and hard limits. Individual container ulimits override these defaults. Valid resource names include: cpu, fsize, data, stack, core, rss, nproc, nofile, memlock, as, locks, sigpending, msgqueue, nice, rtprio, rttime.
@@ -34,6 +68,16 @@ docker:
        readonly: True
        advanced: True
        global: True
+      networks:
+        description: |
+          Docker networks this container is attached to. The first entry is the container's
+          primary network and determines the address its published ports are forwarded to.
+          Defaults to sobridge when unset.
+        helpLink: docker
+        readonly: True
+        advanced: True
+        global: True
+        forcedType: "[]string"
      port_bindings:
        description: List of port bindings for the container.
        helpLink: docker
@@ -30,6 +30,56 @@ fleet_api() {
    curl -sK /opt/so/conf/elasticsearch/curl.config -L "localhost:5601/api/fleet/${QUERYPATH}" "$@" --retry 3 --retry-delay 10 --fail 2>/dev/null
 }

+elastic_fleet_require_agent_policy() {
+    local AGENT_POLICY=$1
+    local POLICY_JSON
+
+    if ! POLICY_JSON=$(fleet_api "agent_policies/$AGENT_POLICY") || [ -z "$POLICY_JSON" ]; then
+        echo "Error: Agent policy '$AGENT_POLICY' was not found or is not visible in the current Kibana space." >&2
+        return 1
+    fi
+
+    if ! jq -e '.item.package_policies | type == "array"' <<<"$POLICY_JSON" >/dev/null 2>&1; then
+        echo "Error: Agent policy '$AGENT_POLICY' was not found or is not visible in the current Kibana space." >&2
+        return 1
+    fi
+
+    echo "$POLICY_JSON"
+}
+
+# Print the single active enrollment token for POLICY_ID.
+# Exit 1: retryable (API failure, invalid response, no active token)
+# Exit 2: multiple active tokens - Shouldn't get into this state without manual intervention
+elastic_fleet_active_enrollment_token() {
+    local POLICY_ID=$1
+    local RESP TOKEN_COUNT API_KEY
+
+    if ! RESP=$(fleet_api "enrollment_api_keys?perPage=100" -H 'kbn-xsrf: true' -H 'Content-Type: application/json'); then
+        echo "Error: Failed to retrieve enrollment tokens for agent policy '$POLICY_ID'." >&2
+        return 1
+    fi
+
+    if ! jq -e '.list' <<<"$RESP" >/dev/null 2>&1; then
+        echo "Error: Invalid enrollment token response for agent policy '$POLICY_ID'." >&2
+        return 1
+    fi
+
+    TOKEN_COUNT=$(jq --arg pid "$POLICY_ID" '[.list[] | select(.policy_id == $pid and .active == true)] | length' <<<"$RESP")
+
+    if [ "${TOKEN_COUNT:-0}" -eq 0 ]; then
+        echo "Error: No active enrollment token found for agent policy '$POLICY_ID'." >&2
+        return 1
+    fi
+
+    if [ "$TOKEN_COUNT" -gt 1 ]; then
+        echo "Error: Found $TOKEN_COUNT active enrollment tokens for agent policy '$POLICY_ID'; expected exactly one." >&2
+        return 2
+    fi
+
+    API_KEY=$(jq -r --arg pid "$POLICY_ID" '.list[] | select(.policy_id == $pid and .active == true) | .api_key' <<<"$RESP")
+    echo "$API_KEY"
+}
+
 # Max number of concurrent Fleet write jobs (create/update). Override via env if needed.
 MAX_FLEET_JOBS=${MAX_FLEET_JOBS:-10}

@@ -62,15 +112,7 @@ elastic_fleet_load_integrations_dir() {
    i=0

    # Fetch the agent policy a single time; we look up integration ids locally below.
-    if ! POLICY_JSON=$(fleet_api "agent_policies/$AGENT_POLICY"); then
-        echo "Error: Failed to retrieve agent policy '$AGENT_POLICY'."
-        rm -f "$FAIL_FILE"
-        rm -rf "$OUT_DIR"
-        return 1
-    fi
-
-    if ! jq -e '.item.package_policies' <<<"$POLICY_JSON" >/dev/null 2>&1; then
-        echo "Error: Invalid agent policy response for '$AGENT_POLICY'."
+    if ! POLICY_JSON=$(elastic_fleet_require_agent_policy "$AGENT_POLICY"); then
        rm -f "$FAIL_FILE"
        rm -rf "$OUT_DIR"
        return 1
@@ -124,9 +166,15 @@ elastic_fleet_integration_check() {

    JSON_STRING=$2

-    NAME=$(jq -r .name $JSON_STRING)
+    NAME=$(jq -r .name "$JSON_STRING")
+    INTEGRATION_ID=""

-    INTEGRATION_ID=$(/usr/sbin/so-elastic-fleet-agent-policy-view "$AGENT_POLICY" | jq -r '.item.package_policies[] | select(.name=="'"$NAME"'") | .id')
+    local POLICY_JSON
+    if ! POLICY_JSON=$(elastic_fleet_require_agent_policy "$AGENT_POLICY"); then
+        return 1
+    fi
+
+    INTEGRATION_ID=$(jq -r --arg n "$NAME" '.item.package_policies[]? | select(.name==$n) | .id' <<<"$POLICY_JSON")

 }

@@ -148,7 +196,16 @@ elastic_fleet_integration_remove() {

    NAME=$2

-    INTEGRATION_ID=$(/usr/sbin/so-elastic-fleet-agent-policy-view "$AGENT_POLICY" | jq -r '.item.package_policies[] | select(.name=="'"$NAME"'") | .id')
+    local POLICY_JSON
+    if ! POLICY_JSON=$(elastic_fleet_require_agent_policy "$AGENT_POLICY"); then
+        return 1
+    fi
+
+    INTEGRATION_ID=$(jq -r --arg n "$NAME" '.item.package_policies[]? | select(.name==$n) | .id' <<<"$POLICY_JSON")
+    if [ -z "$INTEGRATION_ID" ]; then
+        echo "Error: Integration '$NAME' was not found in agent policy '$AGENT_POLICY'." >&2
+        return 1
+    fi

    JSON_STRING=$( jq -n \
                    --arg INTEGRATIONID "$INTEGRATION_ID" \
@@ -13,7 +13,10 @@ ERROR=false
 for INTEGRATION in /opt/so/conf/elastic-fleet/integrations/elastic-defend/*.json
 do
  printf "\n\nInitial Endpoints Policy - Loading $INTEGRATION\n"
-  elastic_fleet_integration_check "endpoints-initial" "$INTEGRATION"
+  if ! elastic_fleet_integration_check "endpoints-initial" "$INTEGRATION"; then
+    ERROR=true
+    continue
+  fi
  if [ -n "$INTEGRATION_ID" ]; then
      printf "\n\nIntegration $NAME exists - Upgrading integration policy\n"
      if ! elastic_fleet_integration_policy_upgrade "$INTEGRATION_ID"; then
@@ -7,20 +7,35 @@
 . /usr/sbin/so-elastic-fleet-common

 # Get all the fleet policies
-json_output=$(curl -s -K /opt/so/conf/elasticsearch/curl.config -L -X GET "localhost:5601/api/fleet/agent_policies" -H 'kbn-xsrf: true')
+if ! json_output=$(fleet_api "agent_policies" -H 'kbn-xsrf: true'); then
+    echo "Error: Failed to retrieve Fleet agent policies." >&2
+    exit 1
+fi
+
+if ! jq -e '.items' <<<"$json_output" >/dev/null 2>&1; then
+    echo "Error: Invalid Fleet agent policies response." >&2
+    exit 1
+fi

 # Extract the IDs that start with "FleetServer_"
-POLICY=$(echo "$json_output" | jq -r '.items[] | select(.id | startswith("FleetServer_")) | .id')
+POLICY=$(jq -r '.items[] | select(.id | startswith("FleetServer_")) | .id' <<<"$json_output")

 # Iterate over each ID in the POLICY variable
 for POLICYNAME in $POLICY; do
    printf "\nUpdating Policy: $POLICYNAME\n"

-    # First get the Integration ID
-    INTEGRATION_ID=$(/usr/sbin/so-elastic-fleet-agent-policy-view "$POLICYNAME" | jq -r '.item.package_policies[] | select(.package.name == "fleet_server") | .id')  
+    if ! POLICY_JSON=$(elastic_fleet_require_agent_policy "$POLICYNAME"); then
+        exit 1
+    fi
+
+    INTEGRATION_ID=$(jq -r '.item.package_policies[]? | select(.package.name == "fleet_server") | .id' <<<"$POLICY_JSON")
+    if [ -z "$INTEGRATION_ID" ]; then
+        echo "Error: fleet_server integration was not found in agent policy '$POLICYNAME'." >&2
+        exit 1
+    fi

    # Modify the default integration policy to update the policy_id and an with the correct naming
-	UPDATED_INTEGRATION_POLICY=$(jq --arg policy_id "$POLICYNAME" --arg name "fleet_server-$POLICYNAME" '
+    UPDATED_INTEGRATION_POLICY=$(jq --arg policy_id "$POLICYNAME" --arg name "fleet_server-$POLICYNAME" '
    .policy_id = $policy_id |
    .name = $name' /opt/so/conf/elastic-fleet/integrations/fleet-server/fleet-server.json)

@@ -22,12 +22,19 @@ NUM_RUNNING=$(pgrep -cf "/bin/bash /sbin/so-elastic-agent-gen-installers")

 for i in {1..30}
 do
-   ENROLLMENTOKEN=$(curl -K /opt/so/conf/elasticsearch/curl.config -L "localhost:5601/api/fleet/enrollment_api_keys?perPage=100" -H 'kbn-xsrf: true' -H 'Content-Type: application/json'  | jq .list | jq -r -c '.[] | select(.policy_id | contains("endpoints-initial")) | .api_key')
+   ENROLLMENTOKEN=$(elastic_fleet_active_enrollment_token "endpoints-initial")
+   TOKEN_RC=$?
+   if [ "$TOKEN_RC" -eq 2 ]; then
+     exit 1
+   fi
   FLEETHOST=$(curl -K /opt/so/conf/elasticsearch/curl.config 'http://localhost:5601/api/fleet/fleet_server_hosts/grid-default' | jq -r  '.item.host_urls[]' | paste -sd ',')
-if [[ $FLEETHOST ]] && [[  $ENROLLMENTOKEN ]]; then break; else sleep 10; fi
+   if [[ -n "$FLEETHOST" ]] && [[ -n "$ENROLLMENTOKEN" ]]; then
+     break
+   fi
+   sleep 10
 done

-if [[ -z $FLEETHOST ]] || [[ -z $ENROLLMENTOKEN ]]; then 
+if [[ -z "$FLEETHOST" ]] || [[ -z "$ENROLLMENTOKEN" ]]; then
  printf "\nFleet Host URL, Enrollment Token or Elastic Version empty - exiting..."
  printf "\nFleet Host: $FLEETHOST, Enrollment Token: $ENROLLMENTOKEN\n"
  exit 1
@@ -67,19 +74,25 @@ for GOOS in "${GOTARGETOS[@]}"; do
    GOARCH="amd64"
    if [[ $GOOS == 'darwin/arm64' ]]; then GOOS="darwin" && GOARCH="arm64"; fi
    printf "\n\n### Generating $GOOS/$GOARCH Installer...\n"
-    docker run -e CGO_ENABLED=0 -e GOOS=$GOOS -e GOARCH=$GOARCH \
+    if ! docker run -e CGO_ENABLED=0 -e GOOS=$GOOS -e GOARCH=$GOARCH \
    --mount type=bind,source=/etc/pki/tls/certs/,target=/workspace/files/cert/ \
    --mount type=bind,source=/nsm/elastic-agent-workspace/,target=/workspace/files/elastic-agent/ \
    --mount type=bind,source=/opt/so/saltstack/local/salt/elasticfleet/files/,target=/output/ \
-    {{ GLOBALS.registry_host }}:5000/{{ GLOBALS.image_repo }}/so-elastic-agent-builder:{{ GLOBALS.so_version }} go build -ldflags "-X main.fleetHostURLsList=$FLEETHOST -X main.enrollmentToken=$ENROLLMENTOKEN"  -o /output/so-elastic-agent_${GOOS}_${GOARCH}
+    {{ GLOBALS.registry_host }}:5000/{{ GLOBALS.image_repo }}/so-elastic-agent-builder:{{ GLOBALS.so_version }} go build -ldflags "-X main.fleetHostURLsList=$FLEETHOST -X main.enrollmentToken=$ENROLLMENTOKEN"  -o /output/so-elastic-agent_${GOOS}_${GOARCH}; then
+        printf "\n### ERROR: Failed to generate $GOOS/$GOARCH installer. Exiting...\n"
+        exit 1
+    fi
    printf "\n### $GOOS/$GOARCH Installer Generated...\n"
 done

 printf "\n\n### Generating MSI...\n"
 cp /opt/so/saltstack/local/salt/elasticfleet/files/so-elastic-agent_windows_amd64 /opt/so/saltstack/local/salt/elasticfleet/files/so-elastic-agent_windows_amd64.exe
-docker run \
+if ! docker run \
 --mount type=bind,source=/opt/so/saltstack/local/salt/elasticfleet/files/,target=/output/ -w /output \
-{{ GLOBALS.registry_host }}:5000/{{ GLOBALS.image_repo }}/so-elastic-agent-builder:{{ GLOBALS.so_version }} wixl -o so-elastic-agent_windows_amd64_msi --arch x64 /workspace/so-elastic-agent.wxs
+{{ GLOBALS.registry_host }}:5000/{{ GLOBALS.image_repo }}/so-elastic-agent-builder:{{ GLOBALS.so_version }} wixl -o so-elastic-agent_windows_amd64_msi --arch x64 /workspace/so-elastic-agent.wxs; then
+    printf "\n### ERROR: Failed to generate MSI. Exiting...\n"
+    exit 1
+fi
 printf "\n### MSI Generated...\n"

 # Verify installers were created
@@ -202,26 +202,9 @@ fi
 ### Finalization ###

 # Query for Enrollment Tokens for default policies
-if ENDPOINTSENROLLMENTOKEN_RAW=$(fleet_api "enrollment_api_keys" -H 'kbn-xsrf: true' -H 'Content-Type: application/json'); then
-    ENDPOINTSENROLLMENTOKEN=$(echo "$ENDPOINTSENROLLMENTOKEN_RAW" | jq .list | jq -r -c '.[] | select(.policy_id | contains("endpoints-initial")) | .api_key')
-else
-    echo -e "\nFailed to query for Endpoints enrollment token"
-    exit 1
-fi
-
-if GRIDNODESENROLLMENTOKENGENERAL_RAW=$(fleet_api "enrollment_api_keys" -H 'kbn-xsrf: true' -H 'Content-Type: application/json'); then
-    GRIDNODESENROLLMENTOKENGENERAL=$(echo "$GRIDNODESENROLLMENTOKENGENERAL_RAW" | jq .list | jq -r -c '.[] | select(.policy_id | contains("so-grid-nodes_general")) | .api_key')
-else
-    echo -e "\nFailed to query for Grid nodes - General enrollment token"
-    exit 1
-fi
-
-if GRIDNODESENROLLMENTOKENHEAVY_RAW=$(fleet_api "enrollment_api_keys" -H 'kbn-xsrf: true' -H 'Content-Type: application/json'); then
-    GRIDNODESENROLLMENTOKENHEAVY=$(echo "$GRIDNODESENROLLMENTOKENHEAVY_RAW" | jq .list | jq -r -c '.[] | select(.policy_id | contains("so-grid-nodes_heavy")) | .api_key')
-else
-    echo -e "\nFailed to query for Grid nodes - Heavy enrollment token"
-    exit 1
-fi
+ENDPOINTSENROLLMENTOKEN=$(elastic_fleet_active_enrollment_token "endpoints-initial") || exit 1
+GRIDNODESENROLLMENTOKENGENERAL=$(elastic_fleet_active_enrollment_token "so-grid-nodes_general") || exit 1
+GRIDNODESENROLLMENTOKENHEAVY=$(elastic_fleet_active_enrollment_token "so-grid-nodes_heavy") || exit 1

 # Store needed data in minion pillar
 pillar_file=/opt/so/saltstack/local/pillar/minions/{{ GLOBALS.minion_id }}.sls
@@ -5,7 +5,8 @@
    { "rename": 	{ "field": "message2.proto", 			"target_field": "network.transport",	"ignore_missing": true 	} },
    { "rename": 	{ "field": "message2.app_proto", 		"target_field": "network.protocol",	"ignore_missing": true 	} },
    { "rename": 	{ "field": "message2.fileinfo.filename", 	"target_field": "file.name",		"ignore_missing": true 	} },
-    { "rename": 	{ "field": "message2.fileinfo.gaps", 		"target_field": "file.bytes.missing",	"ignore_missing": true 	} },
+    { "rename": 	{ "field": "message2.fileinfo.gaps", 		"target_field": "suricata.fileinfo.gaps",	"ignore_missing": true 	} },
+    { "set":		{ "if": "ctx.suricata?.fileinfo?.gaps == false", 	"field": "file.bytes.missing",	"value": 0 		} },
    { "rename": 	{ "field": "message2.fileinfo.magic", 		"target_field": "file.mime_type",	"ignore_missing": true 	} },
    { "rename": 	{ "field": "message2.fileinfo.md5", 		"target_field": "hash.md5",		"ignore_missing": true 	} },
    { "rename": 	{ "field": "message2.fileinfo.sha1", 		"target_field": "hash.sha1",		"ignore_missing": true 	} },
@@ -4,11 +4,19 @@
 {%- set role = GLOBALS.role.split('-')[1] %}
 {%- from 'firewall/containers.map.jinja' import NODE_CONTAINERS %}

+{%- set NODE_NETWORKS = [] %}
+{%- for NETNAME, NETWORK in DOCKERMERGED.networks.items() %}
+{%-   if not NETWORK.get('manager_only') or GLOBALS.get('is_manager', False) %}
+{%-     do NODE_NETWORKS.append(NETNAME) %}
+{%-   endif %}
+{%- endfor %}
+
 {%- set PR = [] %}
 {%- set D1 = [] %}
 {%- set D2 = [] %}
 {%- for container in NODE_CONTAINERS %}
 {%-   set IP = DOCKERMERGED.containers[container].ip %}
+{%-   set BRIDGE = DOCKERMERGED.containers[container].network %}
 {%-   if DOCKERMERGED.containers[container].port_bindings is defined %}
 {%-     for binding in DOCKERMERGED.containers[container].port_bindings %}
 {#-       cant split int so we convert to string #}
@@ -35,11 +43,11 @@
 {%-         endif %}
 {%-         do PR.append("-A POSTROUTING -s " ~ DOCKERMERGED.containers[container].ip ~ "/32 -d " ~ DOCKERMERGED.containers[container].ip ~ "/32 -p " ~  proto ~ " -m " ~ proto  ~ " --dport " ~ containerPort ~ " -j MASQUERADE") %}
 {%-         if bindip | length and bindip != '0.0.0.0' %}
-{%-           do D1.append("-A DOCKER -d " ~ bindip ~ "/32 ! -i sobridge -p " ~ proto ~ " -m " ~ proto ~ " --dport " ~ hostPort ~ " -j DNAT --to-destination " ~ DOCKERMERGED.containers[container].ip ~ ":" ~ containerPort) %}
+{%-           do D1.append("-A DOCKER -d " ~ bindip ~ "/32 ! -i " ~ BRIDGE ~ " -p " ~ proto ~ " -m " ~ proto ~ " --dport " ~ hostPort ~ " -j DNAT --to-destination " ~ DOCKERMERGED.containers[container].ip ~ ":" ~ containerPort) %}
 {%-         else %}
-{%-           do D1.append("-A DOCKER ! -i sobridge -p " ~ proto ~ " -m " ~ proto ~ " --dport " ~ hostPort ~ " -j DNAT --to-destination " ~ DOCKERMERGED.containers[container].ip ~ ":" ~ containerPort) %}
+{%-           do D1.append("-A DOCKER ! -i " ~ BRIDGE ~ " -p " ~ proto ~ " -m " ~ proto ~ " --dport " ~ hostPort ~ " -j DNAT --to-destination " ~ DOCKERMERGED.containers[container].ip ~ ":" ~ containerPort) %}
 {%-         endif %}
-{%-         do D2.append("-A DOCKER -d " ~ DOCKERMERGED.containers[container].ip ~ "/32 ! -i sobridge -o sobridge -p " ~ proto ~ " -m " ~ proto ~ " --dport " ~ containerPort ~ " -j ACCEPT") %}
+{%-         do D2.append("-A DOCKER -d " ~ DOCKERMERGED.containers[container].ip ~ "/32 ! -i " ~ BRIDGE ~ " -o " ~ BRIDGE ~ " -p " ~ proto ~ " -m " ~ proto ~ " --dport " ~ containerPort ~ " -j ACCEPT") %}
 {%-     endfor %}
 {%-   endif %}
 {%- endfor %}
@@ -52,11 +60,15 @@
 :DOCKER - [0:0]
 -A PREROUTING -m addrtype --dst-type LOCAL -j DOCKER
 -A OUTPUT ! -d 127.0.0.0/8 -m addrtype --dst-type LOCAL -j DOCKER
-A POSTROUTING -s {{DOCKERMERGED.range}} ! -o sobridge -j MASQUERADE
+{%- for NETNAME in NODE_NETWORKS %}
+-A POSTROUTING -s {{ DOCKERMERGED.networks[NETNAME].range }} ! -o {{ NETNAME }} -j MASQUERADE
+{%- endfor %}
 {%- for rule in PR %}
 {{ rule }}
 {%- endfor %}
-A DOCKER -i sobridge -j RETURN
+{%- for NETNAME in NODE_NETWORKS %}
+-A DOCKER -i {{ NETNAME }} -j RETURN
+{%- endfor %}
 {%- for rule in D1 %}
 {{ rule }}
 {%- endfor %}
@@ -97,10 +109,12 @@ COMMIT
 {%- endif %}
 -A FORWARD -j DOCKER-USER
 -A FORWARD -j DOCKER-ISOLATION-STAGE-1
-A FORWARD -o sobridge -m conntrack --ctstate RELATED,ESTABLISHED -j ACCEPT
-A FORWARD -o sobridge -j DOCKER
-A FORWARD -i sobridge ! -o sobridge -j ACCEPT
-A FORWARD -i sobridge -o sobridge -j ACCEPT
+{%- for NETNAME in NODE_NETWORKS %}
+-A FORWARD -o {{ NETNAME }} -m conntrack --ctstate RELATED,ESTABLISHED -j ACCEPT
+-A FORWARD -o {{ NETNAME }} -j DOCKER
+-A FORWARD -i {{ NETNAME }} ! -o {{ NETNAME }} -j ACCEPT
+-A FORWARD -i {{ NETNAME }} -o {{ NETNAME }} -j ACCEPT
+{%- endfor %}
 -A FORWARD -m conntrack --ctstate RELATED,ESTABLISHED -j ACCEPT
 -A FORWARD -i lo -j ACCEPT
 -A FORWARD -m conntrack --ctstate INVALID -j DROP
@@ -112,13 +126,18 @@ COMMIT
 {%- for rule in D2 %}
 {{ rule }}
 {%- endfor %}
-
-A DOCKER-ISOLATION-STAGE-1 -i sobridge ! -o sobridge -j DOCKER-ISOLATION-STAGE-2
+{% for NETNAME in NODE_NETWORKS %}
+-A DOCKER-ISOLATION-STAGE-1 -i {{ NETNAME }} ! -o {{ NETNAME }} -j DOCKER-ISOLATION-STAGE-2
+{%- endfor %}
 -A DOCKER-ISOLATION-STAGE-1 -j RETURN
-A DOCKER-ISOLATION-STAGE-2 -o sobridge -j DROP
+{%- for NETNAME in NODE_NETWORKS %}
+-A DOCKER-ISOLATION-STAGE-2 -o {{ NETNAME }} -j DROP
+{%- endfor %}
 -A DOCKER-ISOLATION-STAGE-2 -j RETURN
-A DOCKER-USER ! -i sobridge -o sobridge -m conntrack --ctstate RELATED,ESTABLISHED -j ACCEPT
-A DOCKER-USER ! -i sobridge -o sobridge -j LOGGING
+{%- for NETNAME in NODE_NETWORKS %}
+-A DOCKER-USER ! -i {{ NETNAME }} -o {{ NETNAME }} -m conntrack --ctstate RELATED,ESTABLISHED -j ACCEPT
+-A DOCKER-USER ! -i {{ NETNAME }} -o {{ NETNAME }} -j LOGGING
+{%- endfor %}
 -A DOCKER-USER -j RETURN
 -A LOGGING -m limit --limit 2/min -j LOG --log-prefix "IPTables-dropped: "
 -A LOGGING -j DROP
@@ -4,8 +4,12 @@

 {# add our ip to self #}
 {% do FIREWALL_DEFAULT.firewall.hostgroups.self.append(GLOBALS.node_ip) %}
-{# add dockernet range #}
-{% do FIREWALL_DEFAULT.firewall.hostgroups.dockernet.append(DOCKERMERGED.range) %}
+{# add dockernet ranges #}
+{% for NETNAME, NETWORK in DOCKERMERGED.networks.items() %}
+{%   if not NETWORK.get('manager_only') or GLOBALS.get('is_manager', False) %}
+{%     do FIREWALL_DEFAULT.firewall.hostgroups.dockernet.append(NETWORK.range) %}
+{%   endif %}
+{% endfor %}

 {% if GLOBALS.role == 'so-idh' %}
 {%   from 'idh/opencanary_config.map.jinja' import IDH_PORTGROUPS %}
@@ -26,8 +26,8 @@ so-hydra:
    - hostname: hydra
    - name: so-hydra
    - networks:
-      - sobridge:
-        - ipv4_address: {{ DOCKERMERGED.containers['so-hydra'].ip }}
+      - soauth:
+        - ipv4_address: {{ DOCKERMERGED.containers['so-hydra'].ips['soauth'] }}
    - binds:
      - /opt/so/conf/hydra/:/hydra-conf:ro
      - /opt/so/log/hydra/:/hydra-log:rw
@@ -73,7 +73,7 @@ delete_so-hydra_so-status.disabled:

 wait_for_hydra:
  http.wait_for_successful_query:
-    - name: 'http://{{ GLOBALS.manager }}:4444/health/alive'
+    - name: 'http://{{ DOCKERMERGED.containers['so-hydra'].ips['soauth'] }}:4444/health/alive'
    - ssl: True
    - verify_ssl: False
    - status:
@@ -94,9 +94,11 @@ metrics_link_file:
      - docker_container: so-influxdb

 # Install cron job to determine size of influxdb for telegraf
+# telegraf reads this while the cron rewrites it, so write aside and rename rather than
+# truncating in place. tgraflogdir recurses ownership, so the temp file is chowned to match
 get_influxdb_size:
  cron.present:
-    - name: 'du -s -k /nsm/influxdb | cut -f1 > /opt/so/log/telegraf/influxdb_size.log 2>&1'
+    - name: 'du -s -k /nsm/influxdb | cut -f1 > /opt/so/log/telegraf/influxdb_size.log.tmp 2>&1; chown 939:939 /opt/so/log/telegraf/influxdb_size.log.tmp; mv -f /opt/so/log/telegraf/influxdb_size.log.tmp /opt/so/log/telegraf/influxdb_size.log'
    - identifier: get_influxdb_size
    - user: root
    - minute: '*/1'
@@ -19,8 +19,8 @@ so-kratos:
    - hostname: kratos
    - name: so-kratos
    - networks:
-      - sobridge:
-        - ipv4_address: {{ DOCKERMERGED.containers['so-kratos'].ip }}
+      - soauth:
+        - ipv4_address: {{ DOCKERMERGED.containers['so-kratos'].ips['soauth'] }}
    - binds:
      - /opt/so/conf/kratos/:/kratos-conf:ro
      - /opt/so/log/kratos/:/kratos-log:rw
@@ -71,7 +71,7 @@ delete_so-kratos_so-status.disabled:

 wait_for_kratos:
  http.wait_for_successful_query:
-    - name: 'http://{{ GLOBALS.manager }}:4434/'
+    - name: 'http://{{ DOCKERMERGED.containers['so-kratos'].ips['soauth'] }}:4434/'
    - ssl: True
    - verify_ssl: False
    - status:
@@ -166,7 +166,7 @@ so-repo-sync:

 so_fleetagent_status:
  cron.present:
-    - name: /usr/sbin/so-elasticagent-status > /opt/so/log/agents/agentstatus.log 2>&1
+    - name: '/usr/sbin/so-elasticagent-status > /opt/so/log/agents/agentstatus.log.tmp 2>&1; mv -f /opt/so/log/agents/agentstatus.log.tmp /opt/so/log/agents/agentstatus.log'
    - identifier: so_fleetagent_status
    - user: root
    - minute: '*/5'
@@ -106,7 +106,8 @@ while [[ $# -gt 0 ]]; do
  esac
 done

-hydraUrl=${HYDRA_URL:-http://127.0.0.1:4445}
+hydraContainer=${HYDRA_CONTAINER:-so-hydra}
+hydraUrl=${HYDRA_URL:-http://localhost:4445}
 socRolesFile=${SOC_ROLES_FILE:-/opt/so/conf/soc/soc_clients_roles}
 soUID=${SOCORE_UID:-939}
 soGID=${SOCORE_GID:-939}
@@ -124,6 +125,10 @@ function fail() {
  exit 1
 }

+function hydraCurl() {
+  docker exec "$hydraContainer" curl "$@"
+}
+
 function require() {
  cmd=$1
  which "$1" 2>&1 > /dev/null
@@ -133,8 +138,8 @@ function require() {
 # Verify this environment is capable of running this script
 function verifyEnvironment() {
  require "jq"
-  require "curl"
-  response=$(curl -Ss -L ${hydraUrl}/health/alive)
+  require "docker"
+  response=$(hydraCurl -Ss -L ${hydraUrl}/health/alive)
  [[ "$response" != '{"status":"ok"}' ]] && fail "Unable to communicate with Hydra; specify URL via HYDRA_URL environment variable"
 }

@@ -164,7 +169,7 @@ function ensureRoleFileExists() {
 }

 function listClients() {
-  response=$(curl -Ss -L -f ${hydraUrl}/admin/clients)
+  response=$(hydraCurl -Ss -L -f ${hydraUrl}/admin/clients)
  [[ $? != 0 ]] && fail "Unable to communicate with Hydra"

  clientIds=$(echo "${response}" | jq -r ".[] | .client_id" | sort)
@@ -251,7 +256,7 @@ function createClient() {
 EOF
  )
  
-  response=$(curl -Ss -L --fail-with-body -X POST ${hydraUrl}/admin/clients -d "$body")
+  response=$(hydraCurl -Ss -L --fail-with-body -X POST ${hydraUrl}/admin/clients -d "$body")
  if [[ $? != 0 ]]; then
    error=$(echo $response | jq .error)
    fail "Failed to submit request to Hydra: $error"
@@ -283,7 +288,7 @@ function update() {
 EOF
  )
  
-  response=$(curl -Ss -L --fail-with-body -X PATCH ${hydraUrl}/admin/clients/$id -d "$body")
+  response=$(hydraCurl -Ss -L --fail-with-body -X PATCH ${hydraUrl}/admin/clients/$id -d "$body")
  if [[ $? != 0 ]]; then
    error=$(echo $response | jq .error)
    fail "Failed to submit request to Hydra: $error"
@@ -305,7 +310,7 @@ function generateSecret() {
 EOF
  )
  
-  response=$(curl -Ss -L --fail-with-body -X PATCH ${hydraUrl}/admin/clients/$id -d "$body")
+  response=$(hydraCurl -Ss -L --fail-with-body -X PATCH ${hydraUrl}/admin/clients/$id -d "$body")
  if [[ $? != 0 ]]; then
    error=$(echo $response | jq .error)
    fail "Failed to submit request to Hydra: $error"
@@ -317,7 +322,7 @@ function deleteClient() {

  [[ ${identityId} == "" ]] && fail "Client not found"

-  response=$(curl -Ss -XDELETE -L --fail-with-body "${hydraUrl}/admin/clients/$identityId")
+  response=$(hydraCurl -Ss -XDELETE -L --fail-with-body "${hydraUrl}/admin/clients/$identityId")
  if [[ $? != 0 ]]; then
    error=$(echo $response | jq .error)
    fail "Failed to submit request to Hydra: $error"
@@ -129,7 +129,8 @@ while [[ $# -gt 0 ]]; do
  esac
 done

-kratosUrl=${KRATOS_URL:-http://127.0.0.1:4434/admin}
+kratosContainer=${KRATOS_CONTAINER:-so-kratos}
+kratosUrl=${KRATOS_URL:-http://localhost:4434/admin}
 databasePath=${KRATOS_DB_PATH:-/nsm/kratos/db/db.sqlite}
 databaseTimeout=${KRATOS_DB_TIMEOUT:-5000}
 bcryptRounds=${BCRYPT_ROUNDS:-12}
@@ -154,6 +155,10 @@ function fail() {
  exit 1
 }

+function kratosCurl() {
+  docker exec "$kratosContainer" curl "$@"
+}
+
 function require() {
  cmd=$1
  which "$1" 2>&1 > /dev/null
@@ -164,18 +169,18 @@ function require() {
 function verifyEnvironment() {
  require "htpasswd"
  require "jq"
-  require "curl"
+  require "docker"
  require "openssl"
  require "sqlite3"
  [[ ! -f $databasePath ]] && fail "Unable to find database file; specify path via KRATOS_DB_PATH environment variable"
-  response=$(curl -Ss -L ${kratosUrl}/)
+  response=$(kratosCurl -Ss -L ${kratosUrl}/)
  [[ "$response" != "404 page not found" ]] && fail "Unable to communicate with Kratos; specify URL via KRATOS_URL environment variable"
 }

 function findIdByEmail() {
  email=${1,,}

-  response=$(curl -Ss -L ${kratosUrl}/identities)
+  response=$(kratosCurl -Ss -L ${kratosUrl}/identities)
  identityId=$(echo "${response}" | jq -r ".[] | select(.verifiable_addresses[0].value == \"$email\") | .id")
  echo $identityId
 }
@@ -416,7 +421,7 @@ function syncAll() {
 }

 function listUsers() {
-  response=$(curl -Ss -L ${kratosUrl}/identities)
+  response=$(kratosCurl -Ss -L ${kratosUrl}/identities)
  [[ $? != 0 ]] && fail "Unable to communicate with Kratos"

  users=$(echo "${response}" | jq -r ".[] | .verifiable_addresses[0].value" | sort)
@@ -495,7 +500,7 @@ function createUser() {
 EOF
  )
  
-  response=$(curl -Ss -L ${kratosUrl}/identities -d "$addUserJson")
+  response=$(kratosCurl -Ss -L ${kratosUrl}/identities -d "$addUserJson")
  [[ $? != 0 ]] && fail "Unable to communicate with Kratos"

  identityId=$(echo "${response}" | jq -r ".id")
@@ -518,7 +523,7 @@ function updateStatus() {
  identityId=$(findIdByEmail "$email")
  [[ ${identityId} == "" ]] && fail "User not found"

-  response=$(curl -Ss -L "${kratosUrl}/identities/$identityId")
+  response=$(kratosCurl -Ss -L "${kratosUrl}/identities/$identityId")
  [[ $? != 0 ]] && fail "Unable to communicate with Kratos"

  schemaId=$(echo "$response" | jq -r .schema_id)
@@ -531,7 +536,7 @@ function updateStatus() {
    state="inactive"
  fi
  body="{ \"schema_id\": \"$schemaId\", \"state\": \"$state\", \"traits\": $traitBlock }"
-  response=$(curl -fSsL -XPUT -H "Content-Type: application/json" "${kratosUrl}/identities/$identityId" -d "$body")
+  response=$(kratosCurl -fSsL -XPUT -H "Content-Type: application/json" "${kratosUrl}/identities/$identityId" -d "$body")
  [[ $? != 0 ]] && fail "Unable to update user"
 }

@@ -550,7 +555,7 @@ function updateUserProfile() {
  identityId=$(findIdByEmail "$email")
  [[ ${identityId} == "" ]] && fail "User not found"

-  response=$(curl -Ss -L "${kratosUrl}/identities/$identityId")
+  response=$(kratosCurl -Ss -L "${kratosUrl}/identities/$identityId")
  [[ $? != 0 ]] && fail "Unable to communicate with Kratos"

  schemaId=$(echo "$response" | jq -r .schema_id)
@@ -559,7 +564,7 @@ function updateUserProfile() {
  traitBlock="{\"email\":\"$email\",\"firstName\":\"$firstName\",\"lastName\":\"$lastName\",\"note\":\"$note\"}"

  body="{ \"schema_id\": \"$schemaId\", \"state\": \"$state\", \"traits\": $traitBlock }"
-  response=$(curl -fSsL -XPUT -H "Content-Type: application/json" "${kratosUrl}/identities/$identityId" -d "$body")
+  response=$(kratosCurl -fSsL -XPUT -H "Content-Type: application/json" "${kratosUrl}/identities/$identityId" -d "$body")
  [[ $? != 0 ]] && fail "Unable to update user"
 }

@@ -569,7 +574,7 @@ function deleteUser() {
  identityId=$(findIdByEmail "$email")
  [[ ${identityId} == "" ]] && fail "User not found"

-  response=$(curl -Ss -XDELETE -L "${kratosUrl}/identities/$identityId")
+  response=$(kratosCurl -Ss -XDELETE -L "${kratosUrl}/identities/$identityId")
  [[ $? != 0 ]] && fail "Unable to communicate with Kratos"

  rolesTmpFile="${socRolesFile}.tmp"
@@ -28,6 +28,7 @@ INSTALLEDSALTVERSION=$(salt --versions-report | grep Salt: | awk '{print $2}')
 # percentage like "25%"). Empty means so-soup-grid-highstate uses the salt:auto_apply:batch
 # pillar default.
 BATCHSIZE=
+DEFAULT_DOCKER_RANGE='172.17.1.0/24'
 SOUP_LOG=/root/soup.log
 SOUP_DEBUG_LOG=/root/soup-debug.log
 WHATWOULDYOUSAYYAHDOHERE=soup
@@ -605,6 +606,7 @@ preupgrade_changes() {
    [[ "$INSTALLEDVERSION" == "3.0.0" ]] && up_to_3.1.0
    [[ "$INSTALLEDVERSION" == "3.1.0" ]] && up_to_3.2.0
    [[ "$INSTALLEDVERSION" == "3.2.0" ]] && up_to_3.3.0
+    [[ "$INSTALLEDVERSION" == "3.3.0" ]] && up_to_3.4.0
    true
 }

@@ -623,6 +625,7 @@ postupgrade_changes() {
    [[ "$POSTVERSION" == "3.0.0" ]] && post_to_3.1.0
    [[ "$POSTVERSION" == "3.1.0" ]] && post_to_3.2.0
    [[ "$POSTVERSION" == "3.2.0" ]] && post_to_3.3.0
+    [[ "$POSTVERSION" == "3.3.0" ]] && post_to_3.4.0
    # All applicable post-upgrade steps completed; clear the resume marker.
    rm -f "$POSTVERSION_FILE"
    true
@@ -1173,6 +1176,82 @@ post_to_3.3.0() {
 }
 ### 3.3.0 End ###

+### 3.4.0 Scripts ###
+up_to_3.4.0() {
+  set_soauth_range
+
+  echo "Removing so-kratos, so-hydra and so-soc so they are recreated on the soauth network."
+  docker rm -f so-kratos so-hydra so-soc >> $SOUP_LOG 2>&1
+
+  INSTALLEDVERSION=3.4.0
+}
+
+set_soauth_range() {
+  local pillar_file=/opt/so/saltstack/local/pillar/docker/soc_docker.sls
+  local current_range suggested authnet authgw input
+
+  [[ -f "$pillar_file" ]] || return 0
+
+  current_range=$(so-yaml.py get -r "$pillar_file" docker.range 2>/dev/null) || return 0
+
+  # A default range gets the 172.17.2.0/24 from docker/defaults.yaml, same as a fresh
+  # install, so there is nothing to ask about.
+  [[ -n "$current_range" && "$current_range" != "$DEFAULT_DOCKER_RANGE" ]] || return 0
+
+  if so-yaml.py get -r "$pillar_file" docker.networks.soauth.range >/dev/null 2>&1; then
+    return 0
+  fi
+
+  suggested=$(echo "${current_range%%/*}" | awk -F'.' '{ printf "%s.%s.%s.%s", $1, $2, ($3 + 1) % 256, $4 }')
+
+  if [[ -z $UNATTENDED ]]; then
+    echo ""
+    echo "This grid uses a custom Docker range ($current_range). The authentication"
+    echo "services are moving to their own isolated network, which needs a second /24"
+    echo "that does not overlap it."
+    echo ""
+    while :; do
+      read -rp "Enter the network without the /24 suffix, or press Enter for ${suggested}: " input
+      [[ -z "$input" ]] && input="$suggested"
+      if valid_soauth_range "$input" "$current_range"; then
+        authnet="$input"
+        break
+      fi
+      echo "That range must be a valid IPv4 network, must not be within 172.17.0.0/24, and must not overlap ${current_range}."
+    done
+  else
+    if ! valid_soauth_range "$suggested" "$current_range"; then
+      FINAL_MESSAGE_QUEUE+=("WARNING: Unable to pick a range for the authentication network alongside $current_range. Set it manually before the next highstate:")
+      FINAL_MESSAGE_QUEUE+=("    - so-yaml.py add $pillar_file docker.networks.soauth.range <network>/24")
+      FINAL_MESSAGE_QUEUE+=("    - so-yaml.py add $pillar_file docker.networks.soauth.gateway <gateway>")
+      return 0
+    fi
+    authnet="$suggested"
+    FINAL_MESSAGE_QUEUE+=("NOTE: The authentication services moved to an isolated Docker network and were assigned ${authnet}/24.")
+    FINAL_MESSAGE_QUEUE+=("    - If that conflicts with your environment, update docker.networks.soauth in $pillar_file and run so-checkin.")
+  fi
+
+  authgw=$(echo "$authnet" | awk -F'.' '{print $1,$2,$3,1}' OFS='.')
+
+  echo "Assigning the authentication network the range ${authnet}/24."
+  so-yaml.py add "$pillar_file" docker.networks.soauth.range "${authnet}/24" >> $SOUP_LOG 2>&1
+  so-yaml.py add "$pillar_file" docker.networks.soauth.gateway "$authgw" >> $SOUP_LOG 2>&1
+}
+
+valid_soauth_range() {
+  local candidate=$1 docker_range=$2
+
+  valid_ip4 "$candidate" || return 1
+  [[ $candidate =~ ^172\.17\.0\. ]] && return 1
+  [[ "${candidate}/24" == "$docker_range" ]] && return 1
+  return 0
+}
+
+post_to_3.4.0() {
+  set_postversion 3.4.0
+}
+### 3.4.0 End ###
+

 repo_sync() {
  echo "Sync the local repo."
@@ -1,2 +1,3 @@
-requests>=2.31.0
-whoisit>=2.7.0
+requests>=2.34.0
+whoisit>=4.0.5
+anyio>=4.15.1
@@ -14,6 +14,8 @@
 {%   do SOCDEFAULTS.soc.config.server.modules[module].update({'hostUrl': application_url}) %}
 {% endfor %}

+{% do SOCDEFAULTS.soc.config.server.modules.kratos.update({'publicHostUrl': 'http://' ~ DOCKERMERGED.containers['so-kratos'].ips['soauth'] ~ ':4433/'}) %}
+
 {# add all grid heavy nodes to soc.server.modules.elastic.remoteHostUrls #}
 {% for node_type, minions in salt['pillar.get']('elasticsearch:nodes', {}).items() %}
 {%   if node_type in ['heavynode'] %}
@@ -1380,6 +1380,7 @@ soc:
          retryFailureMaxAttempts: 5
        kratos:
          hostUrl:
+          publicHostUrl:
        hydra:
          hostUrl:
        elastalertengine:
@@ -1465,6 +1466,7 @@ soc:
            - core
            - emerging_threats_addon
          useEsql: false
+          esqlCaseInsensitive: true
        elastic:
          hostUrl:
          remoteHostUrls: []
@@ -1492,6 +1494,9 @@ soc:
          org: Security Onion
          bucket: telegraf/so_short_term
          verifyCert: false
+        notification:
+          dismissedPruneDays: 30
+          enabled: false
        playbook:
          autoUpdateEnabled: true
          playbookImportFrequencySeconds: 86400
@@ -23,7 +23,9 @@ so-soc:
    - name: so-soc
    - networks:
      - sobridge:
-        - ipv4_address: {{ DOCKERMERGED.containers['so-soc'].ip }}
+        - ipv4_address: {{ DOCKERMERGED.containers['so-soc'].ips['sobridge'] }}
+      - soauth:
+        - ipv4_address: {{ DOCKERMERGED.containers['so-soc'].ips['soauth'] }}
    - binds:
      - /nsm/rules:/nsm/rules:rw
      - /opt/so/conf/strelka:/opt/sensoroni/yara:rw
@@ -1,6 +1,31 @@
 name: Security Onion Baseline Pipeline
 priority: 90
 transformations:
+    # ES|QL scalar == returns null on multivalued fields; the
+    # backend reads this key and emits MV_INTERSECTS instead.
+    - id: declare_multivalue_fields
+      type: set_state
+      key: multivalue_fields
+      val:
+        - event.type
+        - event.action
+        - event.category
+        - tags
+        - process.args
+        - related.ip
+        - dns.resolved_ip
+    - id: esql_default_index
+      type: set_state
+      key: index
+      val: .ds-logs-*
+    - id: esql_source_metadata
+      type: set_state
+      key: metadata
+      val: "_id, _index, _source"
+    - id: esql_source_keep
+      type: set_state
+      key: keep
+      val: "_id, _index, _source"
    - id: baseline_field_name_mapping
      type: field_name_mapping
      mapping:
@@ -155,6 +155,14 @@ soc:
        description: Path to custom markdown templates for PDF report generation. All markdown files in this directory will be available as custom reports in the SOC Reports interface.
        global: True
        advanced: True
+      schedules:
+        title: Schedules
+        description: Schedules that are shared across the Security Onion product. Modify via one of the SOC Schedules view.
+        readonlyUi: True
+        global: True
+        forcedType: string
+        syntax: json
+        storage: db
      subgrids:
        title: Subordinate Grids
        description: |
@@ -396,6 +404,11 @@ soc:
            global: True
            advanced: True
            forcedType: bool
+          esqlCaseInsensitive:
+            description: "Match string values case-insensitively when converting Sigma rules. Applies to ES|QL only"
+            global: True
+            advanced: True
+            forcedType: bool
        elastic:
          index:
            description: Comma-separated list of indices or index patterns (wildcard "*" supported) that SOC will search for records.
@@ -476,6 +489,24 @@ soc:
            global: True
            advanced: True
            forcedType: bool
+        notification:
+          destinations:
+            title: Notification Destinations
+            description: JSON list of notifications. Modify via the SOC Notifications view.
+            readonlyUi: True
+            global: True
+            forcedType: string
+            syntax: json
+            storage: db
+          dismissedPruneDays:
+            title: Dismissed Retention Days
+            description: The number of days to retain dismissed notifications. When a notification is dismissed, it will be pruned after this many days. Only one user need dismiss a notification for it to be pruned.
+            forcedType: int
+            global: True
+          enabled:
+            description: Enables or disables the SOC notification module.
+            forcedType: bool
+            global: True
        postgres:
          host:
            description: Hostname or IP address of the PostgreSQL server used by SOC. Defaults to the manager hostname.
@@ -36,7 +36,7 @@ tgraf_sync_script_{{script}}:
    - name: /opt/so/conf/telegraf/scripts/{{script}}
    - user: root
    - group: 939
-    - mode: 770
+    - mode: 750
    - template: jinja
    - source: salt://telegraf/scripts/{{script}}
    - defaults:
@@ -49,7 +49,7 @@ tgraf_sync_script_esindexsize.sh:
    - name: /opt/so/conf/telegraf/scripts/esindexsize.sh
    - user: root
    - group: 939
-    - mode: 770
+    - mode: 750
    - source: salt://telegraf/scripts/esindexsize.sh
 {# Copy conf/elasticsearch/curl.config for telegraf to use with esindexsize.sh #}
 tgraf_sync_escurl_conf:
@@ -61,6 +61,80 @@ tgraf_sync_escurl_conf:
    - source: salt://elasticsearch/curl.config
 {% endif %}

+# so-container-stats runs on the host as somon, a docker group member, so the container does
+# not need the docker socket
+somongroup:
+  group.present:
+    - name: somon
+    - gid: 961
+
+# cron chdirs to $HOME before running a job, so home must exist
+somon:
+  user.present:
+    - uid: 961
+    - gid: 961
+    - home: /opt/so/log/somon
+    - createhome: False
+    - shell: /sbin/nologin
+    - groups:
+      - docker
+    # renumbering an existing somon is a no-op on a fresh host and lets a host created
+    # before the id changed converge instead of failing the whole telegraf state
+    - allow_uid_change: True
+    - allow_gid_change: True
+    - require:
+      - group: somongroup
+
+somonlogdir:
+  file.directory:
+    - name: /opt/so/log/somon
+    - user: 961
+    - group: 961
+    - mode: 755
+    # the lock file is not otherwise managed; recurse so a renumber rechowns it too
+    - recurse:
+      - user
+      - group
+    - require:
+      - user: somon
+
+containers_log:
+  file.managed:
+    - name: /opt/so/log/somon/containers.log
+    - user: 961
+    - group: 961
+    - mode: 644
+    - replace: False
+    - require:
+      - file: somonlogdir
+
+# telegraf reads on the same minute boundary the collector runs, and docker stats takes
+# seconds, so write aside and rename rather than truncating the file telegraf is reading.
+# ; not && so a failed run replaces the file instead of leaving stale metrics behind.
+# flock -n keeps a run that outlives its minute from racing the next one over the same tmp
+# file; the skipped run leaves a stale containers.log, which containers.sh discards by age
+so-container-stats_cron:
+  cron.present:
+    - name: "flock -n /opt/so/log/somon/containers.lock -c '/usr/sbin/so-container-stats > /opt/so/log/somon/containers.log.tmp 2>&1; mv -f /opt/so/log/somon/containers.log.tmp /opt/so/log/somon/containers.log'"
+    - identifier: so-container-stats_cron
+    - user: somon
+    - minute: '*/1'
+    - hour: '*'
+    - daymonth: '*'
+    - month: '*'
+    - dayweek: '*'
+    - require:
+      - user: somon
+
+# salt.lasthighstate touches this at order 9001, after the container starts; pre-create it so
+# docker does not create a directory at the bind mount source
+lasthighstate_placeholder:
+  file.managed:
+    - name: /opt/so/log/salt/lasthighstate
+    - mode: 644
+    - replace: False
+    - makedirs: True
+
 telegraf_sbin:
  file.recurse:
    - name: /usr/sbin
@@ -69,14 +143,20 @@ telegraf_sbin:
    - group: root
    - file_mode: 755

-#telegraf_sbin_jinja:
-#  file.recurse:
-#    - name: /usr/sbin
-#    - source: salt://telegraf/tools/sbin_jinja
-#    - user: 939
-#    - group: 939 
-#    - file_mode: 755
-#    - template: jinja
+# so-container-stats needs the per-stat toggles, so it renders instead of copying
+tgraf_sbin_jinja:
+  file.recurse:
+    - name: /usr/sbin
+    - source: salt://telegraf/tools/sbin_jinja
+    - user: root
+    - group: root
+    - file_mode: 755
+    # the unit test lives beside the script; it must not ship or be rendered as jinja
+    - exclude_pat:
+      - "*_test.py"
+    - template: jinja
+    - defaults:
+        CONTAINER_STATS: {{ TELEGRAFMERGED.container_stats }}

 tgrafconf:
  file.managed:
@@ -10,6 +10,69 @@ telegraf:
    flush_jitter: '0s'
    debug: false
    quiet: false
+  container_stats:
+    tags:
+      identity: False
+    engine:
+      n_containers: False
+      n_containers_running: False
+      n_containers_stopped: False
+      n_containers_paused: False
+      n_images: False
+      n_cpus: False
+      n_goroutines: False
+      n_used_file_descriptors: False
+      n_listener_events: False
+      memory_total: False
+    cpu:
+      usage_percent: True
+      usage_total: False
+      usage_in_usermode: False
+      usage_in_kernelmode: False
+      usage_system: False
+      throttling_periods: False
+      throttling_throttled_periods: False
+      throttling_throttled_time: False
+      container_id: False
+    mem:
+      usage_percent: True
+      usage: False
+      limit: False
+      max_usage: False
+      active_anon: False
+      active_file: False
+      inactive_anon: False
+      inactive_file: False
+      unevictable: False
+      pgfault: False
+      pgmajfault: False
+      container_id: False
+    net:
+      rx_bytes: True
+      rx_packets: False
+      rx_errors: False
+      rx_dropped: False
+      tx_bytes: False
+      tx_packets: False
+      tx_errors: False
+      tx_dropped: False
+      container_id: False
+    blkio:
+      io_service_bytes_recursive_read: False
+      io_service_bytes_recursive_write: False
+      container_id: False
+    status:
+      uptime_ns: True
+      oomkilled: True
+      pid: False
+      exitcode: False
+      restart_count: False
+      started_at: False
+      finished_at: False
+      container_id: False
+    health:
+      health_status: False
+      failing_streak: False
  scripts:
    eval:
      - agentstatus.sh
@@ -19,6 +82,7 @@ telegraf:
      - oldpcap.sh
      - os.sh
      - raid.sh
+      - containers.sh
      - sostatus.sh
      - suriloss.sh
      - surirules.sh
@@ -34,6 +98,7 @@ telegraf:
      - os.sh
      - raid.sh
      - redis.sh
+      - containers.sh
      - sostatus.sh
      - suriloss.sh
      - surirules.sh
@@ -47,6 +112,7 @@ telegraf:
      - os.sh
      - raid.sh
      - redis.sh
+      - containers.sh
      - sostatus.sh
      - features.sh
    managerhype:
@@ -56,6 +122,7 @@ telegraf:
      - os.sh
      - raid.sh
      - redis.sh
+      - containers.sh
      - sostatus.sh
      - features.sh
    managersearch:
@@ -66,12 +133,14 @@ telegraf:
      - os.sh
      - raid.sh
      - redis.sh
+      - containers.sh
      - sostatus.sh
      - features.sh
    import:
      - influxdbsize.sh
      - lasthighstate.sh
      - os.sh
+      - containers.sh
      - sostatus.sh
    sensor:
      - checkfiles.sh
@@ -79,6 +148,7 @@ telegraf:
      - oldpcap.sh
      - os.sh
      - raid.sh
+      - containers.sh
      - sostatus.sh
      - suriloss.sh
      - surirules.sh
@@ -93,6 +163,7 @@ telegraf:
      - os.sh
      - raid.sh
      - redis.sh
+      - containers.sh
      - sostatus.sh
      - suriloss.sh
      - surirules.sh
@@ -101,12 +172,14 @@ telegraf:
    idh:
      - lasthighstate.sh
      - os.sh
+      - containers.sh
      - sostatus.sh
    searchnode:
      - eps.sh
      - lasthighstate.sh
      - os.sh
      - raid.sh
+      - containers.sh
      - sostatus.sh
      - features.sh
    receiver:
@@ -115,16 +188,20 @@ telegraf:
      - os.sh
      - raid.sh
      - redis.sh
+      - containers.sh
      - sostatus.sh
    fleet:
      - lasthighstate.sh
      - os.sh
+      - containers.sh
      - sostatus.sh
    hypervisor:
      - lasthighstate.sh
      - os.sh
+      - containers.sh
      - sostatus.sh
    desktop:
      - lasthighstate.sh
      - os.sh
+      - containers.sh
      - sostatus.sh
@@ -13,6 +13,11 @@ so-telegraf:
  docker_container.absent:
    - force: True

+so-container-stats_cron:
+  cron.absent:
+    - identifier: so-container-stats_cron
+    - user: somon
+
 so-telegraf_so-status.disabled:
  file.comment:
    - name: /opt/so/conf/so-status/so-status.conf
@@ -19,8 +19,7 @@ so-telegraf:
  docker_container.running:
    - image: {{ GLOBALS.registry_host }}:5000/{{ GLOBALS.image_repo }}/so-telegraf:{{ GLOBALS.so_version }}
    - restart_policy: unless-stopped
-    - user: 939
-    - group_add: 939,920
+    - user: 939:939
    - environment:
      - HOST_ETC=/host/etc
      - HOST_SYS=/host/sys
@@ -38,7 +37,6 @@ so-telegraf:
      - /opt/so/conf/telegraf/etc/telegraf.conf:/etc/telegraf/telegraf.conf:ro
      - /opt/so/conf/telegraf/node_config.json:/etc/telegraf/node_config.json:ro
      - /var/run/utmp:/var/run/utmp:ro
-      - /var/run/docker.sock:/var/run/docker.sock:ro
      - /:/host:ro
      - /sys:/host/sys:ro
      - /proc:/host/proc:ro
@@ -51,7 +49,8 @@ so-telegraf:
      - /opt/so/log/suricata:/var/log/suricata:ro
      - /opt/so/log/raid:/var/log/raid:ro
      - /opt/so/log/sostatus:/var/log/sostatus:ro
-      - /opt/so/log/salt:/var/log/salt:ro
+      - /opt/so/log/somon:/var/log/somon:ro
+      - /opt/so/log/salt/lasthighstate:/var/log/salt/lasthighstate:ro
      - /opt/so/log/agents:/var/log/agents:ro
      {% if GLOBALS.is_manager or GLOBALS.role == 'so-heavynode' %}
      - /opt/so/conf/telegraf/etc/escurl.config:/etc/telegraf/elasticsearch.config:ro
@@ -74,6 +73,8 @@ so-telegraf:
    {%   endfor %}
    {% endif %}
    - watch:
+      - file: tgraf_sbin_jinja
+      - file: lasthighstate_placeholder
      - file: trusttheca
      - x509: telegraf_crt
      - x509: telegraf_key
@@ -83,6 +84,8 @@ so-telegraf:
      - file: tgraf_sync_script_{{script}}
    {% endfor %}
    - require:
+      - file: lasthighstate_placeholder
+      - file: somonlogdir
      - file: trusttheca
      - x509: telegraf_crt
      - x509: telegraf_key
@@ -228,13 +228,6 @@
 #   ## bond interfaces.
 #   # bond_interfaces = ["bond0"]

-# # Read metrics about docker containers
- [[inputs.docker]]
-#   ## Docker Endpoint
-#   ##   To use TCP, set endpoint = "tcp://[ip]:[port]"
-#   ##   To use environment variables (ie, docker-machine), set endpoint = "ENV"
-   endpoint = "unix:///var/run/docker.sock"
-#

 # # Read stats from one or more Elasticsearch servers or clusters
 {%- if GLOBALS.is_manager or GLOBALS.role == 'so-heavynode' %}
@@ -342,6 +335,17 @@
   interval = "60s"
 {%- endif %}

+{%- if 'containers.sh' in TELEGRAFMERGED.scripts[GLOBALS.role.split('-')[1]] %}
+{%-   do TELEGRAFMERGED.scripts[GLOBALS.role.split('-')[1]].remove('containers.sh') %}
+[[inputs.exec]]
+   commands = [
+      ["/scripts/containers.sh"]
+   ]
+   data_format = "influx"
+   timeout = "15s"
+   interval = "60s"
+{%- endif %}
+
 {%- if TELEGRAFMERGED.scripts[GLOBALS.role.split('-')[1]] | length > 0 %}
 [[inputs.exec]]
   commands = [
@@ -0,0 +1,29 @@
+#!/bin/bash
+#
+# Copyright Security Onion Solutions LLC and/or licensed to Security Onion Solutions LLC under one
+# or more contributor license agreements. Licensed under the Elastic License 2.0 as shown at
+# https://securityonion.net/license; you may not use this file except in compliance with the
+# Elastic License 2.0.
+
+
+
+# if this script isn't already running
+if [[ ! "`pidof -x $(basename $0) -o %PPID`" ]]; then
+
+    CONTAINERSLOG=/var/log/somon/containers.log
+    # the collector rewrites this every minute; report nothing rather than repeating a stale
+    # file as if it were current, in case a run was skipped or the collector is wedged
+    MAXAGE=150
+
+    if [ -r "$CONTAINERSLOG" ]; then
+        AGE=$(( $(date +%s) - $(stat -c %Y "$CONTAINERSLOG") ))
+        if [ "$AGE" -le "$MAXAGE" ]; then
+            cat $CONTAINERSLOG
+        fi
+    fi
+
+    exit 0
+
+fi
+
+exit 0
@@ -8,10 +8,12 @@
 # if this script isn't already running
 if [[ ! "`pidof -x $(basename $0) -o %PPID`" ]]; then

-    LAST_HIGHSTATE_END=$([ -e "/var/log/salt/lasthighstate" ] && date -r /var/log/salt/lasthighstate +%s || echo 0)
-    NOW=$(date +%s)
-    HIGHSTATE_AGE_SECONDS=$((NOW-LAST_HIGHSTATE_END))
-    echo "salt highstate_age_seconds=$HIGHSTATE_AGE_SECONDS"
+    if [ -r "/var/log/salt/lasthighstate" ]; then
+        LAST_HIGHSTATE_END=$(date -r /var/log/salt/lasthighstate +%s)
+        NOW=$(date +%s)
+        HIGHSTATE_AGE_SECONDS=$((NOW-LAST_HIGHSTATE_END))
+        echo "salt highstate_age_seconds=$HIGHSTATE_AGE_SECONDS"
+    fi

 fi

@@ -53,6 +53,319 @@ telegraf:
      forcedType: bool
      advanced: True
      helpLink: influxdb
+  container_stats:
+    tags:
+      identity:
+        description: Adds the container_image, container_version, engine_host and server_version tags to every container measurement, as the retired inputs.docker plugin did. Increases series cardinality. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+    engine:
+      n_containers:
+        description: Total number of containers known to the Docker engine, running or not. Part of the engine-level docker measurement. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      n_containers_running:
+        description: Number of containers currently running. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      n_containers_stopped:
+        description: Number of containers currently stopped. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      n_containers_paused:
+        description: Number of containers currently paused. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      n_images:
+        description: Number of container images held by the Docker engine. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      n_cpus:
+        description: Number of CPUs the Docker engine reports for this host. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      n_goroutines:
+        description: Number of goroutines inside the Docker daemon. Diagnostic detail for the daemon itself. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      n_used_file_descriptors:
+        description: Number of file descriptors held open by the Docker daemon. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      n_listener_events:
+        description: Number of event listeners subscribed to the Docker daemon. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      memory_total:
+        description: Total physical memory the Docker engine reports for this host, in bytes. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+    cpu:
+      usage_percent:
+        description: Percentage of host CPU consumed by the container. Required by the Container CPU Usage chart on the Security Onion Performance dashboard.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      usage_total:
+        description: Cumulative CPU time consumed by the container, in nanoseconds. Read from the cgroup. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      usage_in_usermode:
+        description: Cumulative CPU time consumed in user mode, in nanoseconds. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      usage_in_kernelmode:
+        description: Cumulative CPU time consumed in kernel mode, in nanoseconds. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      usage_system:
+        description: Cumulative host-wide CPU time, in nanoseconds. Used as the denominator when calculating container CPU percentage. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      throttling_periods:
+        description: Number of CPU enforcement periods the container has seen. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      throttling_throttled_periods:
+        description: Number of periods in which the container was throttled against its CPU limit. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      throttling_throttled_time:
+        description: Total time the container spent throttled, in nanoseconds. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      container_id: &containerid
+        description: Full 64 character container ID, emitted as a field on this measurement. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+    mem:
+      usage_percent:
+        description: Percentage of its memory limit the container is using. Required by the Container Memory Usage chart on the Security Onion Performance dashboard.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      usage:
+        description: Container memory usage in bytes, excluding reclaimable page cache. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      limit:
+        description: Memory limit for the container in bytes. Reports total host memory when the container is unlimited. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      max_usage:
+        description: Peak memory usage for the container in bytes, read from the cgroup peak counter. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      active_anon:
+        description: Anonymous memory on the active LRU list, in bytes. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      active_file:
+        description: Page cache on the active LRU list, in bytes. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      inactive_anon:
+        description: Anonymous memory on the inactive LRU list, in bytes. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      inactive_file:
+        description: Page cache on the inactive LRU list, in bytes. This is the reclaimable cache subtracted from usage. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      unevictable:
+        description: Memory that cannot be reclaimed, in bytes. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      pgfault:
+        description: Cumulative number of page faults taken by the container. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      pgmajfault:
+        description: Cumulative number of major page faults, those requiring disk access. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      container_id: *containerid
+    net:
+      rx_bytes:
+        description: Bytes received by the container across all interfaces except loopback. Required by the Container Traffic - Inbound chart on the Security Onion Performance dashboard.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      rx_packets:
+        description: Packets received by the container. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      rx_errors:
+        description: Receive errors counted on the container interfaces. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      rx_dropped:
+        description: Received packets dropped by the container interfaces. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      tx_bytes:
+        description: Bytes transmitted by the container. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      tx_packets:
+        description: Packets transmitted by the container. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      tx_errors:
+        description: Transmit errors counted on the container interfaces. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      tx_dropped:
+        description: Transmitted packets dropped by the container interfaces. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      container_id: *containerid
+    blkio:
+      io_service_bytes_recursive_read:
+        description: Cumulative bytes read from block devices by the container. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      io_service_bytes_recursive_write:
+        description: Cumulative bytes written to block devices by the container. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      container_id: *containerid
+    status:
+      uptime_ns:
+        description: How long the container has been running, in nanoseconds. Required by the Container Uptime chart on the Security Onion Performance dashboard.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      oomkilled:
+        description: Whether the container was killed by the kernel out-of-memory handler. Required by the Most Recent Container Events table on the Security Onion Performance dashboard.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      pid:
+        description: Host process ID of the container main process. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      exitcode:
+        description: Exit code of the container main process. Meaningful once the container has stopped. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      restart_count:
+        description: Number of times the Docker engine has restarted this container. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      started_at:
+        description: Time the container last started, as a Unix timestamp in nanoseconds. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      finished_at:
+        description: Time the container last exited, as a Unix timestamp in nanoseconds. Absent for a container that has never exited. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      container_id: *containerid
+    health:
+      health_status:
+        description: Result of the container healthcheck, such as healthy, unhealthy or starting. Only emitted for containers that define a healthcheck. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
+      failing_streak:
+        description: Number of consecutive failed healthchecks. Only emitted for containers that define a healthcheck. Defaults to off.
+        forcedType: bool
+        global: True
+        advanced: True
+        helpLink: influxdb
  scripts:
    eval: &telegrafscripts
      description: List of input.exec scripts to run for this node type. The script must be present in salt/telegraf/scripts.
@@ -0,0 +1,398 @@
+#!/usr/bin/env python3
+
+# Copyright Security Onion Solutions LLC and/or licensed to Security Onion Solutions LLC under one
+# or more contributor license agreements. Licensed under the Elastic License 2.0 as shown at
+# https://securityonion.net/license; you may not use this file except in compliance with the
+# Elastic License 2.0.
+
+# Runs from cron as somon, a docker group member, and emits influx line protocol for telegraf to
+# read. This exists so so-telegraf does not need the docker socket; the container reads the output
+# file instead.
+# Measurement/tag/field names match telegraf's inputs.docker plugin because the InfluxDB
+# dashboards query them directly. Which fields are emitted is set per stat in SOC; see
+# telegraf.container_stats in defaults.yaml.
+
+import json
+
+SETTINGS = json.loads('''{{ CONTAINER_STATS | tojson }}''')
+{% raw %}
+import os
+import re
+import subprocess
+import sys
+from datetime import datetime, timezone
+
+CGROUP_ROOT = '/sys/fs/cgroup'
+NANOSEC = 10**9
+# cgroup and docker report cpu time in microseconds; inputs.docker published nanoseconds
+USEC_TO_NSEC = 1000
+
+
+def want(group, field):
+  return bool(SETTINGS.get(group, {}).get(field))
+
+
+def wants_any(group, fields):
+  return any(want(group, field) for field in fields)
+
+
+def docker(args):
+  proc = subprocess.run(['docker'] + args, stdout=subprocess.PIPE, stderr=subprocess.DEVNULL, encoding='utf-8')
+  if proc.returncode != 0:
+    print('Container system error; unable to query docker', file=sys.stderr)
+    sys.exit(1)
+  return proc.stdout
+
+
+def escape_tag(value):
+  return value.replace(',', '\\,').replace(' ', '\\ ').replace('=', '\\=')
+
+
+def quote(value):
+  return '"{}"'.format(str(value).replace('\\', '\\\\').replace('"', '\\"'))
+
+
+def integer(value):
+  return '{}i'.format(int(value))
+
+
+def unsigned(value):
+  # inputs.docker published the cgroup and network counters as uint64; influx stores u and i
+  # as different field types, so matching it keeps historical series readable
+  return '{}u'.format(max(int(value), 0))
+
+
+def read_text(path):
+  try:
+    with open(path) as handle:
+      return handle.read()
+  except OSError:
+    return ''
+
+
+def read_pairs(path):
+  # cgroup files such as memory.stat and cpu.stat are "key value" per line
+  values = {}
+  for line in read_text(path).splitlines():
+    parts = line.split()
+    if len(parts) == 2:
+      try:
+        values[parts[0]] = int(parts[1])
+      except ValueError:
+        pass
+  return values
+
+
+def read_value(path):
+  raw = read_text(path).strip()
+  try:
+    return int(raw)
+  except ValueError:
+    return None
+
+
+def cgroup_path(pid):
+  # "0::/system.slice/docker-<id>.scope" on cgroup v2
+  for line in read_text('/proc/{}/cgroup'.format(pid)).splitlines():
+    parts = line.split(':', 2)
+    if len(parts) == 3 and parts[1] == '':
+      return CGROUP_ROOT + parts[2]
+  return None
+
+
+def host_memory_total():
+  match = re.search(r'MemTotal:\s+(\d+) kB', read_text('/proc/meminfo'))
+  return int(match.group(1)) * 1024 if match else 0
+
+
+def host_cpu_nanoseconds():
+  # inputs.docker's usage_system is the host-wide cpu time the daemon reads from /proc/stat
+  for line in read_text('/proc/stat').splitlines():
+    if line.startswith('cpu '):
+      ticks = sum(int(value) for value in line.split()[1:])
+      return int(ticks * NANOSEC / os.sysconf('SC_CLK_TCK'))
+  return 0
+
+
+def parse_image(image):
+  # mirrors telegraf's internal/docker ParseImage so the tags match what inputs.docker emitted
+  domain = ''
+  remainder = image
+  if '/' in image:
+    head, _, tail = image.partition('/')
+    if '.' in head or ':' in head or head == 'localhost':
+      domain, remainder = head + '/', tail
+  if ':' in remainder:
+    name, _, version = remainder.rpartition(':')
+    return domain + name, version
+  return domain + remainder, 'unknown'
+
+
+def to_nanoseconds(stamp):
+  # docker emits 9 fractional digits; fromisoformat takes at most 6 before python 3.11
+  stamp = stamp.rstrip('Z')
+  if '.' in stamp:
+    whole, _, frac = stamp.partition('.')
+    stamp = whole + '.' + frac[:6]
+  try:
+    parsed = datetime.fromisoformat(stamp).replace(tzinfo=timezone.utc)
+  except ValueError:
+    return None
+  if parsed.year <= 1:
+    return None
+  return int(parsed.timestamp() * NANOSEC)
+
+
+def percent(value):
+  try:
+    return float(value.strip().rstrip('%'))
+  except ValueError:
+    return 0.0
+
+
+def net_counters(pid):
+  # summed across interfaces except loopback, matching the daemon's per-container totals
+  rows = read_text('/proc/{}/net/dev'.format(pid)).splitlines()[2:]
+  if not rows:
+    return None
+  names = ['rx_bytes', 'rx_packets', 'rx_errors', 'rx_dropped']
+  totals = dict.fromkeys(names + ['tx_bytes', 'tx_packets', 'tx_errors', 'tx_dropped'], 0)
+  found = False
+  for row in rows:
+    iface, _, rest = row.partition(':')
+    if iface.strip() == 'lo':
+      continue
+    columns = rest.split()
+    if len(columns) < 12:
+      continue
+    found = True
+    for index, name in enumerate(names):
+      totals[name] += int(columns[index])
+    for index, name in enumerate(['tx_bytes', 'tx_packets', 'tx_errors', 'tx_dropped']):
+      totals[name] += int(columns[8 + index])
+  return totals if found else None
+
+
+def blkio_counters(path):
+  # inputs.docker emitted the device=total row unconditionally, so a container that has done
+  # no block io reports zeros rather than dropping out of the measurement entirely
+  totals = {'io_service_bytes_recursive_read': 0, 'io_service_bytes_recursive_write': 0}
+  rows = read_text(path + '/io.stat').splitlines()
+  for row in rows:
+    for token in row.split()[1:]:
+      key, _, value = token.partition('=')
+      try:
+        if key == 'rbytes':
+          totals['io_service_bytes_recursive_read'] += int(value)
+        elif key == 'wbytes':
+          totals['io_service_bytes_recursive_write'] += int(value)
+      except ValueError:
+        pass
+  return totals
+
+
+def collect_stats():
+  stats = {}
+  for line in docker(['stats', '--no-stream', '--format', '{{json .}}']).splitlines():
+    line = line.strip()
+    if not line:
+      continue
+    try:
+      entry = json.loads(line)
+    except json.JSONDecodeError:
+      continue
+    stats[entry.get('Name', '')] = entry
+  return stats
+
+
+def collect_inspect():
+  ids = docker(['ps', '-aq']).split()
+  if not ids:
+    return []
+  try:
+    return json.loads(docker(['inspect'] + ids))
+  except json.JSONDecodeError:
+    return []
+
+
+def collect_info():
+  try:
+    return json.loads(docker(['info', '--format', '{{json .}}']))
+  except json.JSONDecodeError:
+    return {}
+
+
+class Emitter:
+  def __init__(self):
+    self.lines = []
+
+  def add(self, measurement, tags, fields):
+    if not fields:
+      return
+    tagset = ','.join('{}={}'.format(key, escape_tag(value)) for key, value in tags)
+    body = ','.join('{}={}'.format(key, fields[key]) for key in sorted(fields))
+    self.lines.append('{},{} {}'.format(measurement, tagset, body))
+
+
+def main():
+  cpu_extra = ['usage_total', 'usage_in_usermode', 'usage_in_kernelmode',
+               'throttling_periods', 'throttling_throttled_periods', 'throttling_throttled_time']
+  mem_extra = ['usage', 'limit', 'max_usage', 'active_anon', 'active_file', 'inactive_anon',
+               'inactive_file', 'unevictable', 'pgfault', 'pgmajfault']
+  net_fields = ['rx_bytes', 'rx_packets', 'rx_errors', 'rx_dropped',
+                'tx_bytes', 'tx_packets', 'tx_errors', 'tx_dropped']
+  engine_fields = ['n_containers', 'n_containers_running', 'n_containers_stopped',
+                   'n_containers_paused', 'n_images', 'n_cpus', 'n_goroutines',
+                   'n_used_file_descriptors', 'n_listener_events']
+
+  identity = want('tags', 'identity')
+  need_stats = want('cpu', 'usage_percent')
+  need_cgroup_cpu = wants_any('cpu', cpu_extra)
+  need_cgroup_mem = wants_any('mem', mem_extra + ['usage_percent'])
+  need_blkio = wants_any('blkio', ['io_service_bytes_recursive_read',
+                                   'io_service_bytes_recursive_write', 'container_id'])
+  need_net = wants_any('net', net_fields + ['container_id'])
+  need_info = identity or wants_any('engine', engine_fields + ['memory_total'])
+
+  emitter = Emitter()
+  stats = collect_stats() if need_stats else {}
+  info = collect_info() if need_info else {}
+  engine_tags = [('engine_host', info.get('Name', '')),
+                 ('server_version', info.get('ServerVersion', ''))]
+  system_ns = host_cpu_nanoseconds() if want('cpu', 'usage_system') else 0
+  mem_total = host_memory_total() if wants_any('mem', ['limit', 'usage_percent']) else 0
+
+  if info:
+    counts = {'n_containers': 'Containers', 'n_containers_running': 'ContainersRunning',
+              'n_containers_stopped': 'ContainersStopped', 'n_containers_paused': 'ContainersPaused',
+              'n_images': 'Images', 'n_cpus': 'NCPU', 'n_goroutines': 'NGoroutines',
+              'n_used_file_descriptors': 'NFd', 'n_listener_events': 'NEventsListener'}
+    fields = {name: integer(info[key]) for name, key in counts.items()
+              if want('engine', name) and key in info}
+    emitter.add('docker', engine_tags, fields)
+    # inputs.docker published memory_total as its own point
+    if want('engine', 'memory_total') and 'MemTotal' in info:
+      emitter.add('docker', engine_tags, {'memory_total': integer(info['MemTotal'])})
+
+  for container in collect_inspect():
+    name = container.get('Name', '').lstrip('/')
+    if not name:
+      continue
+    state = container.get('State', {})
+    status = state.get('Status', 'unknown')
+    pid = state.get('Pid')
+    container_id = container.get('Id', '')
+    tags = [('container_name', name), ('container_status', status)]
+    if identity:
+      image, version = parse_image(container.get('Config', {}).get('Image', ''))
+      tags += [('container_image', image), ('container_version', version)] + engine_tags
+
+    status_fields = {}
+    if want('status', 'uptime_ns') or want('status', 'started_at') or want('status', 'finished_at'):
+      started = to_nanoseconds(state.get('StartedAt', ''))
+      finished = to_nanoseconds(state.get('FinishedAt', ''))
+      if started is not None:
+        if want('status', 'started_at'):
+          status_fields['started_at'] = integer(started)
+        if want('status', 'uptime_ns'):
+          end = finished if finished is not None and finished >= started else int(
+            datetime.now(timezone.utc).timestamp() * NANOSEC)
+          status_fields['uptime_ns'] = integer(end - started)
+      if finished is not None and want('status', 'finished_at'):
+        status_fields['finished_at'] = integer(finished)
+    if want('status', 'oomkilled'):
+      status_fields['oomkilled'] = 'true' if state.get('OOMKilled') else 'false'
+    if want('status', 'pid'):
+      status_fields['pid'] = integer(pid or 0)
+    if want('status', 'exitcode'):
+      status_fields['exitcode'] = integer(state.get('ExitCode', 0))
+    if want('status', 'restart_count'):
+      status_fields['restart_count'] = integer(container.get('RestartCount', 0))
+    if want('status', 'container_id'):
+      status_fields['container_id'] = quote(container_id)
+    emitter.add('docker_container_status', tags, status_fields)
+
+    health = state.get('Health')
+    if health:
+      health_fields = {}
+      if want('health', 'health_status'):
+        health_fields['health_status'] = quote(health.get('Status', ''))
+      if want('health', 'failing_streak'):
+        health_fields['failing_streak'] = integer(health.get('FailingStreak', 0))
+      emitter.add('docker_container_health', tags, health_fields)
+
+    if status != 'running' or not pid:
+      continue
+    path = cgroup_path(pid)
+    entry = stats.get(name, {})
+
+    cpu_fields = {}
+    if want('cpu', 'usage_percent'):
+      cpu_fields['usage_percent'] = percent(entry.get('CPUPerc', '0%'))
+    if want('cpu', 'usage_system'):
+      cpu_fields['usage_system'] = unsigned(system_ns)
+    if want('cpu', 'container_id'):
+      cpu_fields['container_id'] = quote(container_id)
+    if need_cgroup_cpu and path:
+      cpu = read_pairs(path + '/cpu.stat')
+      mapping = {'usage_total': 'usage_usec', 'usage_in_usermode': 'user_usec',
+                 'usage_in_kernelmode': 'system_usec',
+                 'throttling_periods': 'nr_periods',
+                 'throttling_throttled_periods': 'nr_throttled',
+                 'throttling_throttled_time': 'throttled_usec'}
+      for field, key in mapping.items():
+        if want('cpu', field) and key in cpu:
+          scale = 1 if field.startswith('throttling_') and field != 'throttling_throttled_time' else USEC_TO_NSEC
+          cpu_fields[field] = unsigned(cpu[key] * scale)
+    emitter.add('docker_container_cpu', tags + [('cpu', 'cpu-total')], cpu_fields)
+
+    mem_fields = {}
+    if want('mem', 'container_id'):
+      mem_fields['container_id'] = quote(container_id)
+    if need_cgroup_mem and path:
+      memory = read_pairs(path + '/memory.stat')
+      for field in ['active_anon', 'active_file', 'inactive_anon', 'inactive_file',
+                    'unevictable', 'pgfault', 'pgmajfault']:
+        if want('mem', field) and field in memory:
+          mem_fields[field] = unsigned(memory[field])
+      current = read_value(path + '/memory.current')
+      # inputs.docker reports usage net of reclaimable page cache
+      usage = max(current - memory.get('inactive_file', 0), 0) if current is not None else None
+      raw_limit = read_value(path + '/memory.max')
+      limit = raw_limit if raw_limit is not None else mem_total
+      if want('mem', 'usage') and usage is not None:
+        mem_fields['usage'] = unsigned(usage)
+      if want('mem', 'limit'):
+        mem_fields['limit'] = unsigned(limit)
+      if want('mem', 'usage_percent') and usage is not None:
+        # same ratio inputs.docker computes, from the same two values
+        mem_fields['usage_percent'] = usage / limit * 100.0 if limit else 0.0
+      if want('mem', 'max_usage'):
+        peak = read_value(path + '/memory.peak')
+        if peak is not None:
+          mem_fields['max_usage'] = unsigned(peak)
+    emitter.add('docker_container_mem', tags, mem_fields)
+
+    # inputs.docker emitted nothing for host-network containers; its Networks map was empty
+    if need_net and container.get('HostConfig', {}).get('NetworkMode', '') != 'host':
+      counters = net_counters(pid)
+      if counters is not None:
+        net_out = {field: unsigned(counters[field]) for field in net_fields if want('net', field)}
+        if want('net', 'container_id'):
+          net_out['container_id'] = quote(container_id)
+        emitter.add('docker_container_net', tags + [('network', 'total')], net_out)
+
+    if need_blkio and path:
+      counters = blkio_counters(path)
+      if counters is not None:
+        blkio_out = {field: unsigned(value) for field, value in counters.items() if want('blkio', field)}
+        if want('blkio', 'container_id'):
+          blkio_out['container_id'] = quote(container_id)
+        emitter.add('docker_container_blkio', tags + [('device', 'total')], blkio_out)
+
+  print('\n'.join(emitter.lines))
+
+
+if __name__ == '__main__':
+  main()
+{% endraw %}
@@ -0,0 +1,635 @@
+# Copyright Security Onion Solutions LLC and/or licensed to Security Onion Solutions LLC under one
+# or more contributor license agreements. Licensed under the Elastic License 2.0 as shown at
+# https://securityonion.net/license; you may not use this file except in compliance with the
+# Elastic License 2.0.
+
+# so-container-stats replaces telegraf's inputs.docker plugin, which was dropped along with the
+# docker socket. The InfluxDB dashboards query its output by measurement, tag and field name, so
+# these tests pin that contract: the shipped defaults, the per-stat toggles, the field types, and
+# the value semantics copied from the plugin.
+#
+# The collector is a jinja template, so every test renders it the way salt does and imports the
+# result. Docker and the cgroup filesystem are faked, so nothing here needs a container runtime.
+
+import contextlib
+import importlib.util
+import io
+import json
+import os
+import tempfile
+import unittest
+
+import jinja2
+import yaml
+
+HERE = os.path.dirname(os.path.abspath(__file__))
+TEMPLATE = os.path.join(HERE, 'so-container-stats')
+DEFAULTS = os.path.join(HERE, '..', '..', 'defaults.yaml')
+
+# a container id is a 64 character hex string; the plugin published it in full
+SOC_ID = 'a' * 64
+TELEGRAF_ID = 'b' * 64
+NGINX_ID = 'c' * 64
+IDSTOOLS_ID = 'd' * 64
+
+
+def shipped_defaults():
+    with open(DEFAULTS) as handle:
+        return yaml.safe_load(handle)['telegraf']['container_stats']
+
+
+def all_enabled():
+    return {group: {field: True for field in fields} for group, fields in shipped_defaults().items()}
+
+
+def render(settings):
+    """Render the template as salt does, import it, and hand back the module."""
+    with open(TEMPLATE) as handle:
+        source = handle.read()
+    rendered = jinja2.Template(source, keep_trailing_newline=True).render(CONTAINER_STATS=settings)
+    path = os.path.join(tempfile.mkdtemp(), 'so_container_stats.py')
+    with open(path, 'w') as handle:
+        handle.write(rendered)
+    spec = importlib.util.spec_from_file_location('so_container_stats', path)
+    module = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(module)
+    return module
+
+
+def split_escaped(text, sep):
+    """Split on an unescaped, unquoted separator, the way influx line protocol is written."""
+    parts, current, escaped, quoted = [], '', False, False
+    for char in text:
+        if escaped:
+            current += char
+            escaped = False
+        elif char == '\\':
+            current += char
+            escaped = True
+        elif char == '"':
+            quoted = not quoted
+            current += char
+        elif char == sep and not quoted:
+            parts.append(current)
+            current = ''
+        else:
+            current += char
+    parts.append(current)
+    return parts
+
+
+def split_on_space(line):
+    escaped = quoted = False
+    for index, char in enumerate(line):
+        if escaped:
+            escaped = False
+        elif char == '\\':
+            escaped = True
+        elif char == '"':
+            quoted = not quoted
+        elif char == ' ' and not quoted:
+            return line[:index], line[index + 1:]
+    return line, ''
+
+
+def parse(output):
+    """Parse line protocol into {(measurement, container): (tags, fields)}."""
+    points = {}
+    for line in output.splitlines():
+        line = line.strip()
+        if not line:
+            continue
+        head, fieldpart = split_on_space(line)
+        pieces = split_escaped(head, ',')
+        measurement, tags = pieces[0], {}
+        for piece in pieces[1:]:
+            key, _, value = piece.partition('=')
+            tags[key] = value
+        fields = {}
+        for piece in split_escaped(fieldpart, ','):
+            key, _, value = piece.partition('=')
+            fields[key] = value
+        key = (measurement, tags.get('container_name', ''))
+        if key in points:
+            # the engine measurement is published as two points, the way inputs.docker did
+            points[key][1].update(fields)
+        else:
+            points[key] = (tags, fields)
+    return points
+
+
+def field_type(value):
+    if value.endswith('u'):
+        return 'unsigned'
+    if value.endswith('i'):
+        return 'integer'
+    if value.startswith('"'):
+        return 'string'
+    if value in ('true', 'false'):
+        return 'boolean'
+    return 'float'
+
+
+def container(name, cid, pid, status='running', network='bridge', image='registry:5000/repo/img:3.4.0',
+              started='2026-09-24T10:00:00.123456789Z', finished='0001-01-01T00:00:00Z', health=None,
+              oomkilled=False, exitcode=0, restarts=0):
+    state = {'Status': status, 'Pid': pid, 'StartedAt': started, 'FinishedAt': finished,
+             'OOMKilled': oomkilled, 'ExitCode': exitcode}
+    if health is not None:
+        state['Health'] = health
+    return {'Id': cid, 'Name': '/' + name, 'State': state, 'RestartCount': restarts,
+            'Config': {'Image': image}, 'HostConfig': {'NetworkMode': network}}
+
+
+class CollectorTestCase(unittest.TestCase):
+    """Builds a fake docker engine and cgroup tree, then runs the collector against it."""
+
+    def setUp(self):
+        self.inspect = [
+            container('so-soc', SOC_ID, 1001),
+            container('so-telegraf', TELEGRAF_ID, 1002, network='host'),
+            container('so-nginx', NGINX_ID, 1003, health={'Status': 'healthy', 'FailingStreak': 0}),
+        ]
+        self.stats = {
+            'so-soc': {'Name': 'so-soc', 'CPUPerc': '1.25%', 'MemPerc': '6.39%', 'NetIO': '1kB / 2kB'},
+            'so-telegraf': {'Name': 'so-telegraf', 'CPUPerc': '0.04%', 'MemPerc': '0.58%', 'NetIO': '0B / 0B'},
+            'so-nginx': {'Name': 'so-nginx', 'CPUPerc': '0.00%', 'MemPerc': '0.09%', 'NetIO': '3kB / 4kB'},
+        }
+        self.info = {'Name': 'sohost', 'ServerVersion': '29.2.1', 'Containers': 3, 'ContainersRunning': 3,
+                     'ContainersStopped': 0, 'ContainersPaused': 0, 'Images': 9, 'NCPU': 8,
+                     'NGoroutines': 42, 'NFd': 77, 'NEventsListener': 1, 'MemTotal': 16000000000}
+        # one cgroup per container, addressed through /proc/<pid>/cgroup exactly as the collector does
+        self.files = {
+            '/proc/meminfo': 'MemTotal:       15625000 kB\n',
+            '/proc/stat': 'cpu  100 200 300 400\ncpu0 1 2 3 4\n',
+        }
+        for pid in (1001, 1002, 1003):
+            self.files['/proc/%d/cgroup' % pid] = '0::/scope%d\n' % pid
+            self.cgroup(pid, 'cpu.stat',
+                        'usage_usec 1000\nuser_usec 600\nsystem_usec 400\nnr_periods 5\nnr_throttled 2\nthrottled_usec 700\n')
+            self.cgroup(pid, 'memory.stat',
+                        'active_anon 300\nactive_file 40\ninactive_anon 200\ninactive_file 50\nunevictable 0\npgfault 1234\npgmajfault 56\n')
+            self.cgroup(pid, 'memory.current', '1000\n')
+            self.cgroup(pid, 'memory.max', '4000\n')
+            self.cgroup(pid, 'memory.peak', '2500\n')
+            self.cgroup(pid, 'io.stat', '8:0 rbytes=100 wbytes=200 rios=1 wios=2\n252:0 rbytes=10 wbytes=20 rios=1 wios=1\n')
+            self.files['/proc/%d/net/dev' % pid] = (
+                'Inter-|   Receive                          |  Transmit\n'
+                ' face |bytes packets errs drop fifo frame compressed multicast|bytes packets errs drop fifo colls carrier compressed\n'
+                '    lo:  9999      99    9    9    0     0          0         0   9999      99    9    9    0     0       0          0\n'
+                '  eth0:  1000      10    1    2    0     0          0         0   2000      20    3    4    0     0       0          0\n'
+                '  eth1:   500       5    0    0    0     0          0         0   1000      10    0    0    0     0       0          0\n')
+
+    def cgroup(self, pid, name, contents):
+        self.files['/sys/fs/cgroup/scope%d/%s' % (pid, name)] = contents
+
+    def run_collector(self, settings=None, module=None):
+        module = module or render(settings if settings is not None else all_enabled())
+
+        def fake_docker(args):
+            if args[0] == 'stats':
+                return ''.join(json.dumps(entry) + '\n' for entry in self.stats.values())
+            if args[0] == 'ps':
+                return ' '.join(entry['Id'] for entry in self.inspect)
+            if args[0] == 'inspect':
+                return json.dumps(self.inspect)
+            if args[0] == 'info':
+                return json.dumps(self.info)
+            raise AssertionError('unexpected docker call: %s' % args)
+
+        module.docker = fake_docker
+        module.read_text = lambda path: self.files.get(path, '')
+        module.CGROUP_ROOT = '/sys/fs/cgroup'
+        buffer = io.StringIO()
+        with contextlib.redirect_stdout(buffer):
+            module.main()
+        self.output = buffer.getvalue()
+        return parse(self.output)
+
+
+class TestShippedDefaults(CollectorTestCase):
+
+    def test_defaults_emit_only_the_dashboard_fields(self):
+        # the Security Onion Performance dashboard queries exactly these five
+        points = self.run_collector(shipped_defaults())
+        emitted = {(measurement, field) for (measurement, _), (_, fields) in points.items() for field in fields}
+        self.assertEqual(emitted, {
+            ('docker_container_cpu', 'usage_percent'),
+            ('docker_container_mem', 'usage_percent'),
+            ('docker_container_net', 'rx_bytes'),
+            ('docker_container_status', 'uptime_ns'),
+            ('docker_container_status', 'oomkilled'),
+        })
+
+    def test_defaults_do_not_emit_the_opt_in_measurements(self):
+        points = self.run_collector(shipped_defaults())
+        measurements = {measurement for measurement, _ in points}
+        self.assertNotIn('docker', measurements)
+        self.assertNotIn('docker_container_blkio', measurements)
+        self.assertNotIn('docker_container_health', measurements)
+
+    def test_defaults_carry_the_tags_the_dashboard_filters_on(self):
+        points = self.run_collector(shipped_defaults())
+        tags, _ = points[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(tags['container_status'], 'running')
+        self.assertEqual(tags['cpu'], 'cpu-total')
+        # identity tags are opt in, so they must be absent by default
+        self.assertNotIn('container_image', tags)
+        self.assertNotIn('engine_host', tags)
+
+
+class TestToggles(CollectorTestCase):
+
+    def test_enabling_one_stat_adds_only_that_field(self):
+        settings = shipped_defaults()
+        settings['cpu']['usage_total'] = True
+        _, fields = self.run_collector(settings)[('docker_container_cpu', 'so-soc')]
+        self.assertIn('usage_total', fields)
+        self.assertNotIn('usage_in_usermode', fields)
+
+    def test_disabling_one_stat_leaves_its_neighbours(self):
+        settings = all_enabled()
+        settings['blkio']['io_service_bytes_recursive_read'] = False
+        _, fields = self.run_collector(settings)[('docker_container_blkio', 'so-soc')]
+        self.assertNotIn('io_service_bytes_recursive_read', fields)
+        self.assertIn('io_service_bytes_recursive_write', fields)
+
+    def test_identity_tags_are_added_when_enabled(self):
+        tags, _ = self.run_collector(all_enabled())[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(tags['container_image'], 'registry:5000/repo/img')
+        self.assertEqual(tags['container_version'], '3.4.0')
+        self.assertEqual(tags['engine_host'], 'sohost')
+        self.assertEqual(tags['server_version'], '29.2.1')
+
+    def test_everything_off_emits_nothing(self):
+        settings = {group: {field: False for field in fields} for group, fields in shipped_defaults().items()}
+        self.assertEqual(self.run_collector(settings), {})
+
+
+class TestFieldTypes(CollectorTestCase):
+    """inputs.docker wrote the cgroup and network counters as unsigned; influx treats u and i as
+    different field types, so a mismatch breaks queries spanning the change."""
+
+    def test_counter_fields_are_unsigned(self):
+        points = self.run_collector(all_enabled())
+        for measurement, field in (('docker_container_cpu', 'usage_total'),
+                                   ('docker_container_cpu', 'throttling_periods'),
+                                   ('docker_container_mem', 'usage'),
+                                   ('docker_container_mem', 'limit'),
+                                   ('docker_container_mem', 'pgfault'),
+                                   ('docker_container_net', 'rx_bytes'),
+                                   ('docker_container_blkio', 'io_service_bytes_recursive_read')):
+            _, fields = points[(measurement, 'so-soc')]
+            self.assertEqual(field_type(fields[field]), 'unsigned', '%s.%s' % (measurement, field))
+
+    def test_status_and_engine_fields_are_signed(self):
+        points = self.run_collector(all_enabled())
+        _, status = points[('docker_container_status', 'so-soc')]
+        for field in ('uptime_ns', 'pid', 'exitcode', 'restart_count', 'started_at'):
+            self.assertEqual(field_type(status[field]), 'integer', field)
+        _, engine = points[('docker', '')]
+        self.assertEqual(field_type(engine['n_containers']), 'integer')
+
+    def test_percentages_are_floats_and_ids_are_quoted_strings(self):
+        points = self.run_collector(all_enabled())
+        _, cpu = points[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(field_type(cpu['usage_percent']), 'float')
+        self.assertEqual(field_type(cpu['container_id']), 'string')
+        self.assertEqual(cpu['container_id'], '"%s"' % SOC_ID)
+        _, status = points[('docker_container_status', 'so-soc')]
+        self.assertEqual(field_type(status['oomkilled']), 'boolean')
+        _, health = points[('docker_container_health', 'so-nginx')]
+        self.assertEqual(health['health_status'], '"healthy"')
+
+
+class TestMemorySemantics(CollectorTestCase):
+
+    def test_usage_subtracts_reclaimable_page_cache(self):
+        # inputs.docker reports usage net of inactive_file: 1000 - 50
+        _, fields = self.run_collector(all_enabled())[('docker_container_mem', 'so-soc')]
+        self.assertEqual(fields['usage'], '950u')
+
+    def test_usage_percent_is_usage_over_limit(self):
+        _, fields = self.run_collector(all_enabled())[('docker_container_mem', 'so-soc')]
+        self.assertAlmostEqual(float(fields['usage_percent']), 950 / 4000 * 100.0)
+
+    def test_limit_falls_back_to_host_memory_when_unlimited(self):
+        for pid in (1001, 1002, 1003):
+            self.cgroup(pid, 'memory.max', 'max\n')
+        _, fields = self.run_collector(all_enabled())[('docker_container_mem', 'so-soc')]
+        self.assertEqual(fields['limit'], '%du' % (15625000 * 1024))
+
+    def test_max_usage_reports_the_cgroup_peak(self):
+        # the daemon reports 0 on cgroup v2, so this deliberately carries the real peak
+        _, fields = self.run_collector(all_enabled())[('docker_container_mem', 'so-soc')]
+        self.assertEqual(fields['max_usage'], '2500u')
+
+
+class TestCpuSemantics(CollectorTestCase):
+
+    def test_microsecond_counters_are_published_as_nanoseconds(self):
+        _, fields = self.run_collector(all_enabled())[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(fields['usage_total'], '1000000u')
+        self.assertEqual(fields['usage_in_usermode'], '600000u')
+        self.assertEqual(fields['usage_in_kernelmode'], '400000u')
+        self.assertEqual(fields['throttling_throttled_time'], '700000u')
+
+    def test_throttling_counts_are_not_scaled(self):
+        _, fields = self.run_collector(all_enabled())[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(fields['throttling_periods'], '5u')
+        self.assertEqual(fields['throttling_throttled_periods'], '2u')
+
+    def test_usage_system_is_host_wide_cpu_time(self):
+        _, fields = self.run_collector(all_enabled())[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(field_type(fields['usage_system']), 'unsigned')
+        self.assertNotEqual(fields['usage_system'], '0u')
+
+
+class TestNetwork(CollectorTestCase):
+
+    def test_counters_sum_interfaces_and_ignore_loopback(self):
+        _, fields = self.run_collector(all_enabled())[('docker_container_net', 'so-soc')]
+        self.assertEqual(fields['rx_bytes'], '1500u')
+        self.assertEqual(fields['rx_packets'], '15u')
+        self.assertEqual(fields['tx_bytes'], '3000u')
+        self.assertEqual(fields['rx_dropped'], '2u')
+
+    def test_host_network_containers_emit_no_row(self):
+        # inputs.docker skipped these: its Networks map is empty for --net=host
+        points = self.run_collector(all_enabled())
+        self.assertNotIn(('docker_container_net', 'so-telegraf'), points)
+        self.assertIn(('docker_container_net', 'so-soc'), points)
+
+    def test_total_tag_is_present(self):
+        tags, _ = self.run_collector(all_enabled())[('docker_container_net', 'so-soc')]
+        self.assertEqual(tags['network'], 'total')
+
+
+class TestBlkio(CollectorTestCase):
+
+    def test_counters_sum_devices(self):
+        _, fields = self.run_collector(all_enabled())[('docker_container_blkio', 'so-soc')]
+        self.assertEqual(fields['io_service_bytes_recursive_read'], '110u')
+        self.assertEqual(fields['io_service_bytes_recursive_write'], '220u')
+
+    def test_container_with_no_io_reports_zero_rather_than_disappearing(self):
+        for pid in (1001, 1002, 1003):
+            self.cgroup(pid, 'io.stat', '')
+        tags, fields = self.run_collector(all_enabled())[('docker_container_blkio', 'so-soc')]
+        self.assertEqual(fields['io_service_bytes_recursive_read'], '0u')
+        self.assertEqual(tags['device'], 'total')
+
+
+class TestStatus(CollectorTestCase):
+
+    def test_running_container_uptime_counts_from_start(self):
+        _, fields = self.run_collector(all_enabled())[('docker_container_status', 'so-soc')]
+        self.assertGreater(int(fields['uptime_ns'].rstrip('i')), 0)
+        self.assertNotIn('finished_at', fields)
+
+    def test_exited_container_reports_its_lifetime_and_finished_at(self):
+        self.inspect.append(container('so-idstools', IDSTOOLS_ID, 0, status='exited',
+                                      started='2026-09-24T10:00:00.000000000Z',
+                                      finished='2026-09-24T10:00:02.000000000Z', exitcode=3, restarts=1))
+        points = self.run_collector(all_enabled())
+        tags, fields = points[('docker_container_status', 'so-idstools')]
+        self.assertEqual(tags['container_status'], 'exited')
+        self.assertEqual(fields['uptime_ns'], '2000000000i')
+        self.assertEqual(fields['finished_at'], '1790244002000000000i')
+        self.assertEqual(fields['exitcode'], '3i')
+        self.assertEqual(fields['restart_count'], '1i')
+        # a stopped container has no live stats, so only the status row is emitted
+        self.assertNotIn(('docker_container_cpu', 'so-idstools'), points)
+
+    def test_health_is_emitted_only_for_containers_with_a_healthcheck(self):
+        points = self.run_collector(all_enabled())
+        self.assertIn(('docker_container_health', 'so-nginx'), points)
+        self.assertNotIn(('docker_container_health', 'so-soc'), points)
+
+    def test_oomkilled_is_reported(self):
+        self.inspect[0]['State']['OOMKilled'] = True
+        _, fields = self.run_collector(all_enabled())[('docker_container_status', 'so-soc')]
+        self.assertEqual(fields['oomkilled'], 'true')
+
+
+class TestEngineMeasurement(CollectorTestCase):
+
+    def test_engine_counts_come_from_docker_info(self):
+        _, fields = self.run_collector(all_enabled())[('docker', '')]
+        self.assertEqual(fields['n_containers'], '3i')
+        self.assertEqual(fields['n_cpus'], '8i')
+        self.assertEqual(fields['n_used_file_descriptors'], '77i')
+
+    def test_engine_is_published_as_two_points(self):
+        # inputs.docker emitted memory_total on its own point, so keep that shape
+        self.run_collector(all_enabled())
+        engine = [line for line in self.output.splitlines() if line.startswith('docker,')]
+        self.assertEqual(len(engine), 2)
+        self.assertTrue(any('memory_total=' in line for line in engine))
+        self.assertTrue(any('n_containers=' in line for line in engine))
+
+    def test_engine_rows_are_tagged_with_host_and_version(self):
+        tags, _ = self.run_collector(all_enabled())[('docker', '')]
+        self.assertEqual(tags['engine_host'], 'sohost')
+        self.assertEqual(tags['server_version'], '29.2.1')
+
+
+class TestLineProtocol(CollectorTestCase):
+
+    def test_tag_values_are_escaped(self):
+        self.inspect[0]['Name'] = '/odd name,with=chars'
+        self.stats['odd name,with=chars'] = self.stats.pop('so-soc')
+        self.stats['odd name,with=chars']['Name'] = 'odd name,with=chars'
+        output = self.run_collector(all_enabled())
+        self.assertIn(('docker_container_cpu', 'odd\\ name\\,with\\=chars'), output)
+
+    def test_image_without_a_tag_reports_version_unknown(self):
+        self.inspect[0]['Config']['Image'] = 'busybox'
+        tags, _ = self.run_collector(all_enabled())[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(tags['container_image'], 'busybox')
+        self.assertEqual(tags['container_version'], 'unknown')
+
+    def test_every_line_has_a_measurement_tagset_and_fieldset(self):
+        module = render(all_enabled())
+
+        def fake_docker(args):
+            if args[0] == 'stats':
+                return ''.join(json.dumps(entry) + '\n' for entry in self.stats.values())
+            if args[0] == 'ps':
+                return ' '.join(entry['Id'] for entry in self.inspect)
+            if args[0] == 'inspect':
+                return json.dumps(self.inspect)
+            return json.dumps(self.info)
+
+        module.docker = fake_docker
+        module.read_text = lambda path: self.files.get(path, '')
+        buffer = io.StringIO()
+        with contextlib.redirect_stdout(buffer):
+            module.main()
+        lines = [line for line in buffer.getvalue().splitlines() if line.strip()]
+        self.assertTrue(lines)
+        for line in lines:
+            head, fieldpart = split_on_space(line)
+            self.assertIn(',', head, line)
+            self.assertIn('=', fieldpart, line)
+            self.assertFalse(fieldpart.endswith(','), line)
+
+
+# group -> (measurement, the container whose row carries it)
+GROUP_TARGET = {
+    'engine': ('docker', ''),
+    'cpu': ('docker_container_cpu', 'so-soc'),
+    'mem': ('docker_container_mem', 'so-soc'),
+    'net': ('docker_container_net', 'so-soc'),
+    'blkio': ('docker_container_blkio', 'so-soc'),
+    'status': ('docker_container_status', 'so-soc'),
+    'health': ('docker_container_health', 'so-nginx'),
+}
+
+
+class TestEverySetting(CollectorTestCase):
+    """Whatever is offered in defaults.yaml has to actually be collectable. These tests are driven
+    off that file, so a new setting that is never wired up fails here rather than shipping."""
+
+    def setUp(self):
+        super().setUp()
+        # an exited container so status.finished_at has a value to report
+        self.inspect.append(container('so-idstools', IDSTOOLS_ID, 0, status='exited',
+                                      started='2026-09-24T10:00:00.000000000Z',
+                                      finished='2026-09-24T10:00:02.000000000Z'))
+
+    def test_every_setting_emits_its_field_when_enabled(self):
+        points = self.run_collector(all_enabled())
+        for group, fields in shipped_defaults().items():
+            if group == 'tags':
+                continue
+            measurement, name = GROUP_TARGET[group]
+            if group == 'status':
+                # finished_at only exists for a container that has actually exited
+                _, exited = points[(measurement, 'so-idstools')]
+                self.assertIn('finished_at', exited)
+            _, emitted = points[(measurement, name)]
+            for field in fields:
+                if group == 'status' and field == 'finished_at':
+                    continue
+                self.assertIn(field, emitted, '%s.%s is offered but never emitted' % (group, field))
+
+    def test_every_setting_is_individually_wired(self):
+        # enabling one stat on its own must produce exactly that field, proving each toggle is
+        # read rather than riding along with a neighbour
+        for group, fields in shipped_defaults().items():
+            if group == 'tags':
+                continue
+            measurement, name = GROUP_TARGET[group]
+            for field in fields:
+                settings = {other: {key: False for key in values} for other, values in shipped_defaults().items()}
+                settings[group][field] = True
+                target = 'so-idstools' if (group == 'status' and field == 'finished_at') else name
+                points = self.run_collector(settings)
+                self.assertIn((measurement, target), points, '%s.%s emitted no row' % (group, field))
+                _, emitted = points[(measurement, target)]
+                self.assertEqual(sorted(emitted), [field], '%s.%s did not emit itself alone' % (group, field))
+
+    def test_all_enabled_values_are_exact(self):
+        points = self.run_collector(all_enabled())
+        clock = os.sysconf('SC_CLK_TCK')
+        expected = {
+            ('docker_container_cpu', 'so-soc'): {
+                'usage_percent': '1.25', 'usage_total': '1000000u', 'usage_in_usermode': '600000u',
+                'usage_in_kernelmode': '400000u', 'usage_system': '%du' % int(1000 * 10**9 / clock),
+                'throttling_periods': '5u', 'throttling_throttled_periods': '2u',
+                'throttling_throttled_time': '700000u', 'container_id': '"%s"' % SOC_ID,
+            },
+            ('docker_container_mem', 'so-soc'): {
+                'usage': '950u', 'limit': '4000u', 'max_usage': '2500u', 'active_anon': '300u',
+                'active_file': '40u', 'inactive_anon': '200u', 'inactive_file': '50u',
+                'unevictable': '0u', 'pgfault': '1234u', 'pgmajfault': '56u',
+                'usage_percent': '23.75', 'container_id': '"%s"' % SOC_ID,
+            },
+            ('docker_container_net', 'so-soc'): {
+                'rx_bytes': '1500u', 'rx_packets': '15u', 'rx_errors': '1u', 'rx_dropped': '2u',
+                'tx_bytes': '3000u', 'tx_packets': '30u', 'tx_errors': '3u', 'tx_dropped': '4u',
+                'container_id': '"%s"' % SOC_ID,
+            },
+            ('docker_container_blkio', 'so-soc'): {
+                'io_service_bytes_recursive_read': '110u',
+                'io_service_bytes_recursive_write': '220u', 'container_id': '"%s"' % SOC_ID,
+            },
+            ('docker', ''): {
+                'n_containers': '3i', 'n_containers_running': '3i', 'n_containers_stopped': '0i',
+                'n_containers_paused': '0i', 'n_images': '9i', 'n_cpus': '8i', 'n_goroutines': '42i',
+                'n_used_file_descriptors': '77i', 'n_listener_events': '1i',
+                'memory_total': '16000000000i',
+            },
+            ('docker_container_health', 'so-nginx'): {
+                'health_status': '"healthy"', 'failing_streak': '0i',
+            },
+        }
+        for key, fields in expected.items():
+            _, emitted = points[key]
+            for field, value in fields.items():
+                self.assertEqual(emitted[field], value, '%s %s' % (key[0], field))
+
+    def test_status_values_are_exact(self):
+        # uptime is relative to now, so it is checked separately from the fixed fields
+        points = self.run_collector(all_enabled())
+        _, running = points[('docker_container_status', 'so-soc')]
+        self.assertEqual(running['pid'], '1001i')
+        self.assertEqual(running['exitcode'], '0i')
+        self.assertEqual(running['restart_count'], '0i')
+        self.assertEqual(running['oomkilled'], 'false')
+        self.assertEqual(running['container_id'], '"%s"' % SOC_ID)
+        self.assertEqual(int(running['started_at'].rstrip('i')) // 10**9, 1790244000)
+        self.assertGreater(int(running['uptime_ns'].rstrip('i')), 0)
+        _, exited = points[('docker_container_status', 'so-idstools')]
+        self.assertEqual(exited['uptime_ns'], '2000000000i')
+        self.assertEqual(exited['finished_at'], '1790244002000000000i')
+
+    def test_tags_identity_toggle_controls_the_identity_tags(self):
+        settings = shipped_defaults()
+        settings['tags']['identity'] = False
+        tags, _ = self.run_collector(settings)[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(sorted(tags), ['container_name', 'container_status', 'cpu'])
+        settings['tags']['identity'] = True
+        tags, _ = self.run_collector(settings)[('docker_container_cpu', 'so-soc')]
+        self.assertEqual(sorted(tags), ['container_image', 'container_name', 'container_status',
+                                        'container_version', 'cpu', 'engine_host', 'server_version'])
+
+    def test_identity_tags_cover_every_documented_tag(self):
+        tags, _ = self.run_collector(all_enabled())[('docker_container_cpu', 'so-soc')]
+        for tag in ('container_image', 'container_version', 'engine_host', 'server_version'):
+            self.assertIn(tag, tags)
+
+
+class TestTemplate(unittest.TestCase):
+
+    def test_template_renders_to_valid_python_for_the_shipped_defaults(self):
+        with open(TEMPLATE) as handle:
+            source = handle.read()
+        rendered = jinja2.Template(source, keep_trailing_newline=True).render(CONTAINER_STATS=shipped_defaults())
+        compile(rendered, 'so-container-stats', 'exec')
+        self.assertNotIn('{%', rendered)
+        # the docker format strings must survive rendering untouched
+        self.assertIn('{{json .}}', rendered)
+
+    def test_every_annotated_setting_exists_in_defaults(self):
+        # SOC reads both trees; an annotation without a default cannot be reverted in the UI
+        with open(os.path.join(HERE, '..', '..', 'soc_telegraf.yaml')) as handle:
+            annotated = yaml.safe_load(handle)['telegraf']['container_stats']
+        defaults = shipped_defaults()
+        for group, fields in annotated.items():
+            self.assertIn(group, defaults)
+            for field in fields:
+                self.assertIn(field, defaults[group], '%s.%s annotated but missing from defaults' % (group, field))
+
+    def test_every_default_setting_is_annotated_for_soc(self):
+        with open(os.path.join(HERE, '..', '..', 'soc_telegraf.yaml')) as handle:
+            annotated = yaml.safe_load(handle)['telegraf']['container_stats']
+        for group, fields in shipped_defaults().items():
+            self.assertIn(group, annotated)
+            for field in fields:
+                self.assertIn(field, annotated[group], '%s.%s missing a SOC annotation' % (group, field))
+
+
+if __name__ == '__main__':
+    unittest.main()
@@ -54,8 +54,8 @@
 {%
  do GLOBALS.update({
    'application_urls': {
-      'hydra': 'http://' ~ GLOBALS.manager ~ ':4445/',
-      'kratos': 'http://' ~ GLOBALS.manager ~ ':4434/',
+      'hydra': 'http://' ~ DOCKERMERGED.containers['so-hydra'].ips['soauth'] ~ ':4445/',
+      'kratos': 'http://' ~ DOCKERMERGED.containers['so-kratos'].ips['soauth'] ~ ':4434/',
      'elastic': 'https://' ~ GLOBALS.manager ~ ':9200/',
      'influxdb': 'https://' ~ GLOBALS.manager ~ ':8086/'
    }
@@ -276,9 +276,20 @@ collect_dockernet() {
 			whiptail_invalid_input
 			whiptail_dockernet_sosnet "$DOCKERNET"
 		done
+
+		whiptail_authnet_sosnet "$(adjacent_net "$DOCKERNET")"
+
+		while ! valid_ip4 "$AUTHNET" || [[ $AUTHNET =~ "172.17.0." ]] || [[ "$AUTHNET" == "$DOCKERNET" ]]; do
+			whiptail_invalid_input
+			whiptail_authnet_sosnet "$AUTHNET"
+		done
 	fi
 }

+adjacent_net() {
+	echo "$1" | awk -F'.' '{ printf "%s.%s.%s.%s", $1, $2, ($3 + 1) % 256, $4 }'
+}
+
 collect_gateway() {
 	whiptail_management_interface_gateway

@@ -1399,6 +1410,15 @@ docker_pillar() {
 			"docker:"\
 			"  range: '$DOCKERNET/24'"\
 			"  gateway: '$DOCKERGATEWAY'" > $docker_pillar_file
+
+		if [ ! -z "$AUTHNET" ]; then
+			AUTHGATEWAY=$(echo $AUTHNET | awk -F'.' '{print $1,$2,$3,1}' OFS='.')
+			printf '%s\n'\
+				"  networks:"\
+				"    soauth:"\
+				"      range: '$AUTHNET/24'"\
+				"      gateway: '$AUTHGATEWAY'" >> $docker_pillar_file
+		fi
 	fi
 }

@@ -1591,6 +1611,7 @@ reserve_group_ids() {
 	logCmd "groupadd -g 949 elastic-agent"
 	logCmd "groupadd -g 947 elastic-fleet"
 	logCmd "groupadd -g 960 kafka"
+	logCmd "groupadd -g 961 somon"
 }

 reserve_ports() {
@@ -365,6 +365,18 @@ whiptail_dockernet_sosnet() {

 }

+whiptail_authnet_sosnet() {
+
+	[ -n "$TESTING" ] && return
+
+	AUTHNET=$(whiptail --title "$whiptail_title" --inputbox \
+	"\nEnter a second /24 size network range WITHOUT the /24 suffix. The authentication services are isolated on their own network so that the identity provider is not reachable from other containers. It must not overlap the range you just entered, and any range within 172.17.0.0/24 cannot be used." 13 65 "$1" 3>&1 1>&2 2>&3)
+
+	local exitstatus=$?
+	whiptail_check_exitstatus $exitstatus
+
+}
+
 whiptail_end_settings() {
 	[ -n "$TESTING" ] && return

@@ -427,6 +439,7 @@ whiptail_end_settings() {
 	[[ -n $WEBUSER ]] && __append_end_msg "Web User: $WEBUSER"

 	[[ -n $DOCKERNET ]] && __append_end_msg "Docker network: $DOCKERNET/24"
+	[[ -n $AUTHNET ]] && __append_end_msg "Authentication network: $AUTHNET/24"
 	if [[ ${#ntp_servers[@]} -gt 0 ]]; then
 		__append_end_msg "NTP Servers:"
 		for server in "${ntp_servers[@]}"; do
Author	SHA1	Message	Date
Josh Patterson	72f60fcaa9	FIX: remove the docker socket from so-telegraf so-telegraf mounted /var/run/docker.sock and joined the host docker group on every node type. The :ro flag blocks write() to the inode, not connect() plus HTTP over the socket, so any code execution inside the container could reach POST /containers/create with Privileged:true and become root on the host. No telegraf script used the socket; the only consumer was the native [[inputs.docker]] plugin, and group_add 920 existed solely to feed it. Container metrics now come from so-container-stats, a collector that runs on the host from cron and writes influx line protocol to a file telegraf already had mounted. This is the pattern so-status, so-raid-status and so-elasticagent-status already use, so the privileged docker access stays on the host side where root cron already ran it. The collector runs as somon, a service account in the docker group with no login shell and a locked password, rather than root. Docker group membership is still root-equivalent on the host, so this is defense in depth rather than a privilege boundary. The telegraf scripts were root:939 mode 770, letting socore rewrite them for code execution inside the container; they are now 750, which still allows the read and execute telegraf needs. The container also ran with no group, giving it gid 0, and now runs as 939:939. That alone would have broken lasthighstate.sh, which reached /opt/so/log/salt only via the root group and could not tell an unreadable file from a missing one, so it silently reported a 56 year highstate age. Only the lasthighstate file is bind mounted now, and the script tests readability instead of existence. Everything inputs.docker collected beyond the five fields the shipped dashboards query is available per stat under telegraf:container_stats, annotated for SOC so an operator can enable it without editing files. Defaults reproduce the previous output exactly. With every stat enabled the emitted field set matches what inputs.docker wrote, verified by running the plugin against the live socket and diffing: 53 fields, no type mismatches, no field present on one side only. Two deliberate differences: max_usage carries the real cgroup peak where the daemon reports 0 on cgroup v2, and host-network containers emit no docker_container_net row, matching inputs.docker. Docker label tags are not restored, since nothing queries them and they cost significant cardinality. so-status, the influxdb size cron, so-elasticagent-status, so-raid-status and so-common-status-check truncated their output in place while telegraf read it, so telegraf periodically saw an empty file and logged a parse error or emitted empty values. They now write aside and rename. Measured on a live manager, the old so-status cron left status.log empty for 215 of 10997 reads. Tested on a fresh install, a converted grid and a 3.0 upgrade.	2026-09-24 14:30:25 -04:00
Jorge Reyes	2684a5ca95	Merge pull request #16260 from Security-Onion-Solutions/reyesj2/16254 FIX: Fleet scripts failing when endpoints-initial is missing	2026-09-23 08:58:15 -05:00
reyesj2	bcee63bde5	remove policy precheck	2026-09-23 08:51:11 -05:00
reyesj2	6ce89eb323	jq -e exits 0 with empty input	2026-09-22 21:25:50 -05:00
reyesj2	ebab4b0d90	split between missing token and multiple enrollment tokens	2026-09-22 16:11:56 -05:00
reyesj2	db60c27da2	fail cleanly when endpoints-initial is missing or has multiple enrollment tokens	2026-09-22 12:24:23 -05:00
Jason Ertel	f4defdfde0	Merge pull request #16258 from Security-Onion-Solutions/jertel/wip upgrade whoislookup analyzer deps; notification annotations	2026-09-22 08:41:04 -04:00
Jason Ertel	36652e8f23	fixed malformed desc annotation	2026-09-22 08:30:44 -04:00
Jason Ertel	7bef194540	force transitive dep version	2026-09-22 08:28:44 -04:00
Jason Ertel	e2bf2837fe	notification annotations	2026-09-22 08:18:59 -04:00
Jason Ertel	06704dad22	upgrade whoislookup analyzer deps	2026-09-22 08:17:18 -04:00
Mike Reeves	47fe0758d0	Merge pull request #16255 from Security-Onion-Solutions/mreeves/fix-so-user-stdin-drain Fix user creation failing with a valid password	2026-09-18 17:12:09 -04:00
Mike Reeves	b71fd93f9d	Stop docker exec from consuming so-user's piped password Adding a user failed with "Password does not meet the minimum requirements" for passwords that were well over the eight character minimum. so-user reads the password from stdin in updatePassword, but verifyEnvironment runs first and calls kratosCurl, which ran docker exec with -i. That attaches stdin and drains the pipe the caller sent the password on, so the later read -rs saw EOF, password was empty, and expr length "" tripped the minimum length check. No kratosCurl or hydraCurl call sends a body on stdin; every one passes it as a -d argument, so -i was never needed. Dropping it leaves the admin API responses unchanged. so-client had the same wrapper for the Hydra admin API. It does not currently read stdin, but the flag drains its caller's pipe just the same, so it is dropped there too.	2026-09-18 17:04:17 -04:00
Mike Reeves	d9eff9aa9e	Merge pull request #16253 from Security-Onion-Solutions/mreeves/fix-docker-networks-setting-conflict Fix config tree error from the docker.networks setting	2026-09-18 15:41:25 -04:00
Mike Reeves	d8884dbd99	Document sobridge alongside the soauth network settings sobridge carries most containers but held no annotation, so it was absent from the config tree entirely. Its pillar entry is an empty map, so annotating it adds a setting with no children for docker.networks.soauth.range to collide with, and the tree still builds. docker.map.jinja fills in sobridge's range and gateway at render time from docker.range and docker.gateway, and resets the entry when it is not a mapping, so nothing writes keys underneath it. The sentence about sobridge is dropped from the soauth range description now that sobridge documents itself.	2026-09-18 14:37:11 -04:00
Mike Reeves	6c0d4c15e8	Annotate the soauth network leaves instead of the networks map SOC reported "Malformed config setting (docker.networks.soauth.range): Setting name 'networks' conflicts with another similarly named setting" and refused to render the config tree. soc_docker.yaml annotated docker.networks itself, and every key in that block was a scalar, so FlattenAnnotations registered docker.networks as a setting. defaults.yaml holds a nested map there, so FlattenPillar separately registered docker.networks.soauth.range, .gateway and .manager_only, and HydrateAnnotations unions the two sets. The config tree gives each id segment either a value or children, never both, so addToNode pushed docker.networks as a childless leaf and then threw when docker.networks.soauth.range tried to descend through it. The annotations move down to the three leaves that actually carry values, so docker.networks is a plain branch. This matches docker.containers, which is nested the same way and has never carried annotations of its own. docker.ulimits and the per-container networks list are unaffected: their pillar values are lists rather than maps, so they stay single settings.	2026-09-18 13:42:41 -04:00
Jason Ertel	2e2f62f265	Merge pull request #16252 from Security-Onion-Solutions/jertel/wip store in db	2026-09-18 12:47:32 -04:00
Jason Ertel	b7a11a525c	store in db	2026-09-18 12:41:35 -04:00
Mike Reeves	8eef95ea3e	Merge pull request #16221 from Security-Onion-Solutions/mreeves/kratos-soauth-network Isolate the Kratos admin API on a dedicated soauth docker network	2026-09-18 10:24:00 -04:00
Mike Reeves	65e261475d	Mention Hydra in the soauth network description The description was written alongside the Kratos change and named only the Kratos admin API, but so-hydra now sits on soauth too. The trailing note about rebuilding the firewall is dropped, since the setting is readonly and cannot be changed from SOC in the first place.	2026-09-18 10:17:20 -04:00
Mike Reeves	bbc28c88b7	Isolate the Hydra admin API on the soauth docker network The Hydra admin API creates OAuth clients and introspects tokens without authentication, relying on network isolation the same way Kratos does, but so-hydra published 0.0.0.0:4445:4445 and sat on sobridge. With userland-proxy left at its default of true, Docker ran a proxy listener on the host, so any container reached the admin API through the manager IP or the bridge gateway regardless of its docker network. so-hydra now sits alone on soauth and no longer publishes 4445. 4444 is still published, so the nginx proxy for /oauth2/token and the well-known endpoints is unchanged. so-soc is already dual homed from the Kratos change and reaches the admin API over soauth, so only its hostUrl moves. wait_for_hydra polls the container address instead of the host port. so-client reaches the admin API through docker exec, mirroring so-user. Exit codes still propagate, so the --fail-with-body error handling is unchanged. so-hydra is removed in up_to_3.4.0 alongside so-kratos and so-soc so the highstate recreates it with the correct network membership. The auth network range is already handled by set_soauth_range.	2026-09-17 16:58:49 -04:00
Jason Ertel	edaacf79a7	Merge pull request #16251 from Security-Onion-Solutions/jertel/wip notification annotations	2026-09-17 15:37:04 -04:00
Jason Ertel	ecc643cd33	fix typo	2026-09-17 15:34:29 -04:00
Jason Ertel	08aaf7948e	Merge branch '3/dev' into jertel/wip	2026-09-17 15:32:35 -04:00
Jason Ertel	bb57545d08	new annotations for notifications	2026-09-17 15:32:21 -04:00
Josh Brower	0f7adbbecc	Merge pull request #16247 from Security-Onion-Solutions/esql-fixes Refactor for ESQL	2026-09-17 09:56:19 -04:00
Josh Patterson	aeb4fe8f50	Merge pull request #16250 from Security-Onion-Solutions/fix/root-own-sbin-and-salt-tree FIX: root-own /usr/sbin management scripts and the Salt default tree	2026-09-17 09:44:47 -04:00
defensivedepth	f3aa39c5a4	Tweak name	2026-09-17 09:43:44 -04:00
Doug Burks	24077ba974	Merge pull request #16249 from Security-Onion-Solutions/dougburks/fix-eval-suricata-file-drops FIX: suricata.fileinfo maps boolean gaps into long file.bytes.missing	2026-09-17 09:42:27 -04:00
Doug Burks	b3567405f9	FIX: suricata.fileinfo maps boolean gaps into long file.bytes.missing	2026-09-17 07:57:44 -04:00
defensivedepth	1fc5bb7afa	Refactor for ESQL	2026-09-17 07:56:42 -04:00
Mike Reeves	cf3a4ebc27	Isolate the Kratos admin API on a dedicated soauth docker network The Kratos admin API is unauthenticated by design and relies on network isolation, but so-kratos published 0.0.0.0:4434:4434 and daemon.json leaves userland-proxy at its default of true. Docker therefore ran a proxy listener on the host, and any container reached the admin API through the manager IP or the bridge gateway regardless of its docker network. so-kratos now sits alone on a new soauth network and no longer publishes 4434. 4433 is still published, so nginx is unchanged. so-soc is dual homed and reaches the admin API over soauth, with publicHostUrl set explicitly. so-user reaches it through docker exec, and wait_for_kratos polls the container address instead of the host port. firewall/iptables.jinja hardcoded sobridge in every generated DNAT, ACCEPT, masquerade and isolation rule, so it is now driven by a docker:networks map and a per-container networks list. Containers that do not declare one default to sobridge, and .ip still resolves to the primary network's address. soauth is only created on the roles that run so-kratos. Setup already allows a custom docker range, so it now also prompts for the auth network range and writes it to the docker pillar. Grids on the default range need no decision during soup, since they get the same 172.17.2.0/24 a fresh install would. Grids that set a custom range are prompted, defaulting to the adjacent /24, and unattended upgrades take that default with a notice rather than blocking. so-kratos and so-soc are removed in up_to_3.4.0 so the highstate recreates them with the correct network membership.	2026-09-09 16:09:00 -04:00