chore: publish 2.3.1+v1.21.3 release

I updated to the latest monitoring-ng version, and got the following error:

```
FATA failed to create config stats_something_gf_alerts_traefik_: Error response from daemon: rpc error: code = InvalidArgument desc = invalid name, only 64 [a-zA-Z0-9-_.] characters allowed, and the start and end character must be [a-zA-Z0-9]
```

It seems that abra.sh is missing an export of a new variable.

Reviewed-on: #57
Reviewed-by: moritz <98+moritz@noreply.git.coopcloud.tech>
Co-authored-by: Linus Gasser <ineiti@gasser.blue>

.env.sample

@@ -53,6 +53,13 @@ SECRET_BASIC_AUTH_VERSION=v1
 # Not needed if you just want to read local log files — use SYSLOG_FILES instead.
 # SYSLOG=1
 # COMPOSE_FILE="$COMPOSE_FILE:compose.syslog.yml"
+#
+# Scrape Docker-Events via official docker-cli image using docker socket
+# COMPOSE_FILE="$COMPOSE_FILE:compose.docker-events.yml"
+
+
+# Monitor physical disks health
+# COMPOSE_FILE="$COMPOSE_FILE:compose.smartctl.yml"
 
 # Monitoring Server
 #
@@ -68,10 +75,18 @@ SECRET_BASIC_AUTH_VERSION=v1
 #
 # COMPOSE_FILE="$COMPOSE_FILE:compose.loki.yml"
 #
+# LOKI_LOG_LEVEL=warn
+#
 # Set to 0 to disable retention
 # LOKI_RETENTION_PERIOD=744h
 # LOKI_STORAGE_FILESYSTEM=1
 #
+# Opt in to migrating the deprecated boltdb-shipper index store to TSDB, from this date
+# onward (YYYY-MM-DD). MUST be a FUTURE date, e.g. the day after you deploy this: a past
+# or same-day date makes both stores claim that day's data, breaking queries for it.
+# Logs from before this date are still ingested and queried from boltdb-shipper.
+# LOKI_TSDB_FROM_DATE=2026-11-01
+#
 ## S3 Storage
 # LOKI_STORAGE_S3=1
 # LOKI_AWS_ENDPOINT=https://minio.autonomic.zone
@@ -126,3 +141,15 @@ SECRET_BASIC_AUTH_VERSION=v1
 
 # Node memory usage alert will trigger when memory usage is above the given number in percent
 #ALERT_NODE_MEMORY_USAGE=85
+
+# Loki ingestion alert will trigger when a host ships fewer than the given number of log entries per second
+#ALERT_LOKI_INGESTION_STALLED=0.01
+
+# Loki write error alert will trigger when log pushes fail at more than the given rate per second
+#ALERT_LOKI_WRITE_ERRORS=0
+
+# Loki WAL disk alert will trigger when the filesystem holding the WAL is above the given number in percent. Loki throttles all writes at 90
+#ALERT_LOKI_WAL_DISK_USAGE=85
+
+# Traefik apdex alert will trigger when the apdex score of a service drops below the given number. Services receiving fewer than 10 requests per 10m are ignored
+#ALERT_TRAEFIK_APDEX=0.5

.gitea/CODEOWNERS

@@ -0,0 +1 @@
+* @coop-cloud/monitoring-ng-maintainers

.gitignore

@@ -1 +1,2 @@
 /.envrc
+*~

MAINTENANCE.md

@@ -0,0 +1,35 @@
+# Monitoring NG Recipe Maintenance
+
+All contributions should be made via a pull request. This is to ensure a
+certain quality and consistency, that others can rely on.
+
+## Maintainer Responsibilities
+
+A recipe maintainer has the following responsibilities:
+
+- Respond to pull requests / issues within two weeks
+- Make image security updates within a week
+- Make image major updates every three months
+
+In order to fullfill these responsibilities a recipe maintainer:
+
+- Has to watch the repository (to get notifications)
+- Needs to make sure renovate is configured properly
+
+## Pull Requests
+
+A pull request can be merged if it is approved by at least one maintainer. For
+pull requests opened by a maintainer they need to be approved by another
+maintainer. Even though it is okay to merge a pull request with one approval, it
+is always better if all maintainers looked at the pull request and approved it.
+
+## Become a maintainer
+
+Everyone can apply to be a recipe maintainer:
+1. Watch the repository to always get updates
+2. Simply add your self to the list in the [README.md](./README.md) and open a new pull request with the change.
+3. Once the pull request gets merged you will be added to the
+   [monitoring-ng maintainers
+   team](https://git.coopcloud.tech/org/coop-cloud/teams/monitoring-ng-maintainers).
+4. Join the room
+   [#cc-|-monitoring-ng-maintenance:matrix.org](#cc-|-monitoring-ng-maintenance:matrix.org) and chat to other maintainers.

README.md

@@ -7,7 +7,7 @@ It's based heavily on the [monitoring-lite](https://git.coopcloud.tech/coop-clou
 Metrics and logs are collected by [Grafana Alloy](https://grafana.com/docs/alloy/latest/) and pushed to a central Prometheus/Loki (via `remote_write`/`loki push`). Every `monitoring-ng` instance runs its own Alloy, whether or not it also runs the central stack.
 
 <!-- metadata -->
-
+- **Maintainer**: [@coop-cloud/monitoring-ng-maintainers](https://git.coopcloud.tech/org/coop-cloud/teams/monitoring-ng-maintainers)
 - **Category**: Apps
 - **Status**: 2, beta
 - **Image**: [`grafana/grafana`](https://hub.docker.com/r/grafana/grafana), 4, upstream
@@ -54,6 +54,14 @@ This is what a gathering host pushes into. It also runs its own Alloy, so it mon
 
 ## Additional features
 
+### Scrape docker events
+
+Just uncomment the following line in your .env to collect docker events using
+the official docker-cli image and the docker socket:
+```
+COMPOSE_FILE="$COMPOSE_FILE:compose.docker-events.yml"
+```
+
 ### Discovering metrics from other apps
 
 Alloy auto-discovers and scrapes other Docker Swarm services running on the same host, on the `proxy` network, that opt in via labels. No manual scrape config needed. On the app's `compose.yml`:
@@ -122,6 +130,18 @@ $ echo '{
 $ systemctl restart docker.service
 ```
 
+### Upgrading: Loki TSDB migration
+
+Loki's index store defaults to `boltdb-shipper`, which is deprecated upstream in favour of
+TSDB. To migrate an existing deployment without losing access to already-ingested logs, set
+`LOKI_TSDB_FROM_DATE` in your `.env` (see `.env.sample`) to a date on or after the day after
+your deploy, then `abra app deploy`. Logs from before that date stay on boltdb-shipper and
+remain queryable as before; only logs from that date onward are indexed with TSDB.
+
+**The date must be in the future when you deploy it, never today or earlier**: both stores share the same
+per-day table, so a past or same-day date makes them collide over that day's data and
+breaks querying it.
+
 ### Setup Push Gateway
 
 1. Enable in the env file by uncommenting the following lines:
@@ -166,3 +186,12 @@ It is possible to enable the following alerts, by uncommenting the corresponding
 
 - node disk space: `ALERT_NODE_DISK_SPACE_LEFT`
 - node memory usage: `ALERT_NODE_MEMORY_USAGE`
+- traefik apdex: `ALERT_TRAEFIK_APDEX`
+
+## smart monitoring
+
+To be able monitor hard drive health data, you need to configure
+`smartd` to run on the host system, and also the
+`collect-smartctl-json.sh` script provided here (via cronjob or as
+a `smartd` hook).  This is a limitation on Docker Swarm, which prevents
+the `smartctl_exporter` from running on privileged mode.

abra.sh

@@ -5,12 +5,14 @@ export GF_STACKS_DASH_VERSION=v3
 export GF_TRAEFIK_DASH_VERSION=v3
 export GF_BACKUP_DASH_VERSION=v1
 export GF_CUSTOM_INI_VERSION=v4
-export LOKI_YML_VERSION=v3
+export LOKI_YML_VERSION=v4
 export PROMETHEUS_YML_VERSION=v2
 export MATRIX_ALERTMANAGER_CONFIG_VERSION=v1
 export MATRIX_ALERTMANAGER_ENTRYPOINT_VERSION=v1
-export GF_ALERTS_NODE_VERSION=v3
-export CONFIG_ALLOY_VERSION=v2
+export GF_ALERTS_NODE_VERSION=v5
+export GF_ALERTS_LOKI_VERSION=v5
+export CONFIG_ALLOY_VERSION=v3
+export GF_ALERTS_TRAEFIK_VERSION=v1
 
 # migrates secrets from old names to new names by reading values from the
 # running containers on the server and re-inserting them under the new names.

alerts/loki.yml.tmpl

@@ -0,0 +1,191 @@
+apiVersion: 1
+
+# List of alert rule UIDs that should be deleted
+deleteRules:
+  {{ if not (env "ALERT_LOKI_INGESTION_STALLED") }}
+  - orgId: 1
+    uid: coopcloud_loki_ingestion_stalled
+  {{ end }}
+  {{ if not (env "ALERT_LOKI_WRITE_ERRORS") }}
+  - orgId: 1
+    uid: coopcloud_loki_write_errors
+  {{ end }}
+  {{ if not (env "ALERT_LOKI_WAL_DISK_USAGE") }}
+  - orgId: 1
+    uid: coopcloud_loki_wal_disk_usage
+  {{ end }}
+
+groups:
+- orgId: 1
+  name: loki
+  folder: monitoring
+  interval: 5m
+  rules:
+    {{ if (env "ALERT_LOKI_INGESTION_STALLED") }}
+    - uid: coopcloud_loki_ingestion_stalled
+      title: Loki Ingestion Stalled
+      condition: C
+      data:
+        - refId: A
+          relativeTimeRange:
+            from: 1800
+            to: 0
+          datasourceUid: PBFA97CFB590B2093
+          model:
+            editorMode: code
+            expr: sum by (instance) (rate(loki_write_sent_entries_total[15m]))
+            instant: true
+            intervalMs: 1000
+            legendFormat: __auto
+            maxDataPoints: 43200
+            range: false
+            refId: A
+        - refId: C
+          relativeTimeRange:
+            from: 1800
+            to: 0
+          datasourceUid: __expr__
+          model:
+            conditions:
+                - evaluator:
+                    params:
+                        - {{ env "ALERT_LOKI_INGESTION_STALLED" }}
+                    type: lt
+                  operator:
+                    type: and
+                  query:
+                    params:
+                        - C
+                  reducer:
+                    params: []
+                    type: last
+                  type: query
+            datasource:
+                type: __expr__
+                uid: __expr__
+            expression: A
+            intervalMs: 1000
+            maxDataPoints: 43200
+            refId: C
+            type: threshold
+      noDataState: Alerting
+      execErrState: Error
+      for: 15m
+      annotations:
+        description: ""
+        runbook_url: ""
+        summary: Loki received no logs from {{`{{ $labels.instance }}`}} for 15m ({{`{{ printf "%.4f" (index $values "A").Value }}`}} entries/sec, threshold {{ env "ALERT_LOKI_INGESTION_STALLED" }})
+      isPaused: false
+    {{ end }}
+    {{ if (env "ALERT_LOKI_WRITE_ERRORS") }}
+    - uid: coopcloud_loki_write_errors
+      title: Loki Write Errors
+      condition: C
+      data:
+        - refId: A
+          relativeTimeRange:
+            from: 1200
+            to: 0
+          datasourceUid: PBFA97CFB590B2093
+          model:
+            editorMode: code
+            expr: sum by (instance, status_code) (rate(loki_write_request_duration_seconds_count{status_code!~"2.."}[10m]))
+            instant: true
+            intervalMs: 1000
+            legendFormat: __auto
+            maxDataPoints: 43200
+            range: false
+            refId: A
+        - refId: C
+          relativeTimeRange:
+            from: 1200
+            to: 0
+          datasourceUid: __expr__
+          model:
+            conditions:
+                - evaluator:
+                    params:
+                        - {{ env "ALERT_LOKI_WRITE_ERRORS" }}
+                    type: gt
+                  operator:
+                    type: and
+                  query:
+                    params:
+                        - C
+                  reducer:
+                    params: []
+                    type: last
+                  type: query
+            datasource:
+                type: __expr__
+                uid: __expr__
+            expression: A
+            intervalMs: 1000
+            maxDataPoints: 43200
+            refId: C
+            type: threshold
+      noDataState: NoData
+      execErrState: Error
+      for: 10m
+      annotations:
+        description: ""
+        runbook_url: ""
+        summary: Loki is rejecting log pushes from {{`{{ $labels.instance }}`}} with HTTP {{`{{ $labels.status_code }}`}} ({{`{{ printf "%.2f" (index $values "A").Value }}`}} failed pushes/sec)
+      isPaused: false
+    {{ end }}
+    {{ if (env "ALERT_LOKI_WAL_DISK_USAGE") }}
+    - uid: coopcloud_loki_wal_disk_usage
+      title: Loki WAL Disk Usage
+      condition: C
+      data:
+        - refId: A
+          relativeTimeRange:
+            from: 600
+            to: 0
+          datasourceUid: PBFA97CFB590B2093
+          model:
+            editorMode: code
+            expr: max by (instance) (loki_ingester_wal_disk_usage_percent * 100)
+            instant: true
+            intervalMs: 1000
+            legendFormat: __auto
+            maxDataPoints: 43200
+            range: false
+            refId: A
+        - refId: C
+          relativeTimeRange:
+            from: 600
+            to: 0
+          datasourceUid: __expr__
+          model:
+            conditions:
+                - evaluator:
+                    params:
+                        - {{ env "ALERT_LOKI_WAL_DISK_USAGE" }}
+                    type: gt
+                  operator:
+                    type: and
+                  query:
+                    params:
+                        - C
+                  reducer:
+                    params: []
+                    type: last
+                  type: query
+            datasource:
+                type: __expr__
+                uid: __expr__
+            expression: A
+            intervalMs: 1000
+            maxDataPoints: 43200
+            refId: C
+            type: threshold
+      noDataState: NoData
+      execErrState: Error
+      for: 10m
+      annotations:
+        description: ""
+        runbook_url: ""
+        summary: Loki WAL disk {{`{{ printf "%.1f" (index $values "A").Value }}`}}% full on {{`{{ $labels.instance }}`}} (threshold {{ env "ALERT_LOKI_WAL_DISK_USAGE" }}%, writes stop at 90%)
+      isPaused: false
+    {{ end }}

alerts/node.yml.tmpl

@@ -71,8 +71,7 @@ groups:
         description: ""
         runbook_url: ""
         summary: Less than {{ env "ALERT_NODE_DISK_SPACE_LEFT" }}% disk space left on {{`{{ $labels.instance }}`}} ({{`{{ (index $values "A").Value }}`}}% left)
-      labels:
-        "": ""
+      labels: {}
       isPaused: false
     {{ end }}
     {{ if (env "ALERT_NODE_MEMORY_USAGE")  }}
@@ -87,7 +86,7 @@ groups:
           datasourceUid: PBFA97CFB590B2093
           model:
             editorMode: code
-            expr: (node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100
+            expr: 100 - ((node_memory_MemAvailable_bytes / node_memory_MemTotal_bytes) * 100)
             instant: true
             intervalMs: 1000
             legendFormat: __auto

alerts/traefik.yml.tmpl

@@ -0,0 +1,81 @@
+apiVersion: 1
+
+# List of alert rule UIDs that should be deleted
+deleteRules:
+  {{ if not (env "ALERT_TRAEFIK_APDEX") }}
+  - orgId: 1
+    uid: coopcloud_traefik_apdex
+  {{ end }}
+
+groups:
+- orgId: 1
+  name: traefik
+  folder: recipes
+  interval: 5m
+  rules:
+    {{ if (env "ALERT_TRAEFIK_APDEX") }}
+    - uid: coopcloud_traefik_apdex
+      title: Apdex
+      condition: B
+      data:
+        - refId: A
+          relativeTimeRange:
+            from: 600
+            to: 0
+          datasourceUid: PBFA97CFB590B2093
+          model:
+            editorMode: code
+            # apdex = (satisfied + tolerating / 2) / total, which reduces to
+            # (le=0.3 + le=1.2) / 2 / total because the le=1.2 bucket is
+            # cumulative and therefore already contains the satisfied requests.
+            # services with less than 10 requests per 10m are ignored.
+            expr: |-
+              ((sum(rate(traefik_service_request_duration_seconds_bucket{le="0.3",code="200"}[5m])) by (service) +
+               sum(rate(traefik_service_request_duration_seconds_bucket{le="1.2",code="200"}[5m])) by (service)) / 2 /
+               sum(rate(traefik_service_request_duration_seconds_count{code="200"}[5m])) by (service))
+              and on(service)
+              (
+                sum(increase(traefik_service_request_duration_seconds_count{code="200"}[10m])) by (service) >= 10
+              )
+            instant: true
+            intervalMs: 1000
+            legendFormat: __auto
+            maxDataPoints: 43200
+            range: false
+            refId: A
+        - refId: B
+          relativeTimeRange:
+            from: 600
+            to: 0
+          datasourceUid: __expr__
+          model:
+            conditions:
+                - evaluator:
+                    params:
+                        - {{ env "ALERT_TRAEFIK_APDEX" }}
+                    type: lt
+                  operator:
+                    type: and
+                  query:
+                    params: []
+                  reducer:
+                    params: []
+                    type: avg
+                  type: query
+            datasource:
+                type: __expr__
+                uid: __expr__
+            expression: A
+            intervalMs: 1000
+            maxDataPoints: 43200
+            refId: B
+            type: threshold
+      noDataState: OK
+      execErrState: Error
+      for: 10m
+      annotations:
+        description: ""
+        runbook_url: ""
+        summary: Apdex for {{`{{ $labels.service }}`}} is below {{ env "ALERT_TRAEFIK_APDEX" }} (currently at {{`{{ printf "%.2f" (index $values "A").Value }}`}})
+      isPaused: false
+    {{ end }}

collect-smartctl-json.service

@@ -0,0 +1,6 @@
+[Unit]
+Description=Collect SMART data
+
+[Service]
+Type=oneshot
+ExecStart=/usr/local/bin/collect-smartctl-json.sh

collect-smartctl-json.sh

@@ -0,0 +1,61 @@
+#! /bin/bash
+# Adapted from https://github.com/prometheus-community/smartctl_exporter/blob/master/collect-smartctl-json.sh
+
+# Data directory to dump smartctl output
+# This directory will be created if it doesn't exist
+data_dir="/var/lib/smartmontools/json"
+
+# The original script used --xall but that doesn't work
+# This matches the command in readSMARTctl()
+smartctl_args="--json --info --health --attributes --tolerance=verypermissive \
+--nocheck=standby --format=brief --log=error"
+
+# Ignore this devices
+smartctl_ignore_dev_regex="^(/dev/bus)"
+
+# Determine the json query tool to use
+if command -v jq >/dev/null; then
+	json_tool="jq"
+	json_args="--raw-output"
+elif command -v yq >/dev/null; then
+	json_tool="yq"
+	json_args="--unwrapScalar"
+else
+	echo -e "One of 'yq' or 'jq' is required. Please try again after \
+installing one of them"
+	exit 1
+fi
+
+if [[ ! "${UID}" -eq 0 ]] && ! command -v sudo >/dev/null; then
+	# Not root and sudo doesn't exist
+	echo "sudo does not exist. Please run this as root"
+	exit 1
+fi
+
+SUDO="sudo"
+if [[ "${UID}" -eq 0 ]]; then
+	# Don't use sudo if root
+	SUDO=""
+fi
+
+[[ ! -d "${data_dir}" ]] && mkdir --parents "${data_dir}"
+
+if [[ $# -ne 0 ]]; then
+	devices="${1}"
+else
+	devices="$(smartctl --scan --json | "${json_tool}" "${json_args}" \
+".devices[].name | select(test(\"${smartctl_ignore_dev_regex}\") | not)")"
+  mapfile -t devices <<< "${devices[@]}"
+fi
+
+for device in "${devices[@]}"
+  do
+	echo -n "Collecting data for '${device}'..."
+	# shellcheck disable=SC2086
+	data="$($SUDO smartctl ${smartctl_args} ${device})"
+	# Accommodate a smartmontools pre-7.3 bug
+	data=${data#"  Pending defect count:"}
+	device_name="$(basename "${device}")"
+	echo -e "\tSaving to ${device_name}.json"
+	echo "${data}" > "${data_dir}/${device_name}.json"
+done

collect-smartctl-json.timer

@@ -0,0 +1,9 @@
+[Unit]
+Description=Collect SMART data
+
+[Timer]
+OnCalendar=hourly
+Persistent=true
+
+[Install]
+WantedBy=timers.target

compose.docker-events.yml

@@ -0,0 +1,20 @@
+version: "3.8"
+services:
+  docker-events:
+    image: docker:29.8.1-cli
+    command: >
+      docker events
+      --format '{{json .}}'
+      --filter type=container
+      --filter event=start
+      --filter event=die
+      --filter event=kill
+      --filter event=oom
+      --filter event=restart
+      --filter event=stop
+      --filter event=health_status
+    volumes:
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+    deploy:
+      restart_policy:
+        condition: on-failure

compose.grafana.yml

@@ -2,7 +2,7 @@ version: '3.8'
 
 services:
   grafana:
-    image: grafana/grafana:13.0.6
+    image: grafana/grafana:13.3.0
     volumes:
       - grafana-data:/var/lib/grafana:rw
     secrets:
@@ -24,6 +24,10 @@ services:
         target: /var/lib/grafana/dashboards/backup.json
       - source: gf_alerts_node
         target: /etc/grafana/provisioning/alerting/node.yml
+      - source: gf_alerts_loki
+        target: /etc/grafana/provisioning/alerting/loki.yml
+      - source: gf_alerts_traefik
+        target: /etc/grafana/provisioning/alerting/traefik.yml
     networks:
       - proxy
       - internal
@@ -34,6 +38,10 @@ services:
       - GF_INSTALL_PLUGINS
       - ALERT_NODE_DISK_SPACE_LEFT
       - ALERT_NODE_MEMORY_USAGE
+      - ALERT_LOKI_INGESTION_STALLED
+      - ALERT_LOKI_WRITE_ERRORS
+      - ALERT_LOKI_WAL_DISK_USAGE
+      - ALERT_TRAEFIK_APDEX
     deploy:
       labels:
         - "traefik.enable=true"
@@ -48,7 +56,7 @@ services:
       interval: 5s
       timeout: 10s
       retries: 3
-      start_period: 10s
+      start_period: 300s
 
 configs:
   gf_custom_ini:
@@ -77,6 +85,14 @@ configs:
     template_driver: golang
     name: ${STACK_NAME}_gf_alerts_node_${GF_ALERTS_NODE_VERSION}
     file: alerts/node.yml.tmpl
+  gf_alerts_loki:
+    template_driver: golang
+    name: ${STACK_NAME}_gf_alerts_loki_${GF_ALERTS_LOKI_VERSION}
+    file: alerts/loki.yml.tmpl
+  gf_alerts_traefik:
+    template_driver: golang
+    name: ${STACK_NAME}_gf_alerts_traefik_${GF_ALERTS_TRAEFIK_VERSION}
+    file: alerts/traefik.yml.tmpl
 
 volumes:
   grafana-data:

compose.loki.yml

@@ -2,7 +2,7 @@ version: '3.8'
 
 services:
   loki:
-    image: grafana/loki:3.7.2
+    image: grafana/loki:3.7.9
     command: -config.file=/etc/loki/local-config.yaml
     networks:
       - proxy
@@ -19,9 +19,17 @@ services:
       - LOKI_AWS_REGION
       - LOKI_BUCKET_NAMES
       - LOKI_RETENTION_PERIOD
+      - LOKI_LOG_LEVEL
       - STACK_NAME
       - LOKI_STORAGE_FILESYSTEM
       - LOKI_STORAGE_S3
+      - LOKI_TSDB_FROM_DATE
+    healthcheck:
+      test: ["CMD", "/usr/bin/loki", "-health"]
+      interval: 10s
+      timeout: 10s
+      retries: 10
+      start_period: 30s
     deploy:
       restart_policy:
         condition: on-failure
@@ -34,6 +42,8 @@ services:
         - "traefik.http.routers.${STACK_NAME}-loki.tls=true"
         - "traefik.http.routers.${STACK_NAME}-loki.tls.certresolver=${LETS_ENCRYPT_ENV}"
         - "traefik.http.routers.${STACK_NAME}-loki.middlewares=basicauth@file"
+        - "prometheus.io/scrape=true"
+        - "prometheus.io/port=3100"
 
 
 configs:

compose.prometheus.yml

@@ -2,7 +2,7 @@ version: '3.8'
 
 services:
   prometheus:
-    image: prom/prometheus:v3.12.0
+    image: prom/prometheus:v3.15.1
     secrets:
       - basic_auth
     volumes:
@@ -28,7 +28,7 @@ services:
       start_period: 30s
     deploy:
       restart_policy:
-        condition: on-failure
+        condition: any
       labels:
         - "traefik.enable=true"
         - "traefik.swarm.network=proxy"

compose.smartctl.yml

@@ -0,0 +1,17 @@
+---
+version: "3.8"
+services:
+  smartctl:
+    image: "prometheuscommunity/smartctl-exporter:v0.15.0"
+    volumes:
+    - "/var/lib/smartmontools/json:/debug"
+    command:
+    - "--smartctl.fake-data"
+    - "--smartctl.interval=1h"
+    networks:
+    - "proxy"
+    deploy:
+      labels:
+        - "prometheus.io/scrape=true"
+        - "prometheus.io/port=9633"
+        - "prometheus.io/path=/metrics"

compose.yml

@@ -3,7 +3,7 @@ version: "3.8"
 
 services:
   app:
-    image: grafana/alloy:v1.18.1
+    image: grafana/alloy:v1.21.3
     hostname: "${DOMAIN}"
     configs:
       - source: config_alloy
@@ -46,14 +46,14 @@ services:
       test: ["CMD", "bash", "-c", "exec 3<>/dev/tcp/localhost/12345 && printf 'GET /-/ready HTTP/1.0\r\nHost: localhost\r\n\r\n' >&3 && head -1 <&3 | grep -q 200"]
       interval: 5s
       timeout: 10s
-      retries: 3
-      start_period: 10s
+      retries: 10
+      start_period: 30s
     deploy:
       restart_policy:
-        condition: on-failure
+        condition: any
       labels:
         - "backupbot.backup=${ENABLE_BACKUPS:-true}"
-        - "coop-cloud.${STACK_NAME}.version=1.6.0+v1.8.1"
+        - "coop-cloud.${STACK_NAME}.version=2.3.1+v1.21.3"
 configs:
   config_alloy:
     template_driver: golang

config.alloy.tmpl

@@ -52,7 +52,32 @@ prometheus.scrape "default" {
     prometheus.exporter.cadvisor.docker.targets,
   )
 
+  forward_to = [prometheus.relabel.container_meta.receiver]
+}
+
+prometheus.relabel "container_meta" {
   forward_to = [prometheus.remote_write.prometheus.receiver]
+
+  // remove sha tail: nginx:1.31.1@sha256:608a... -> nginx:1.31.1
+  rule {
+    source_labels = ["image"]
+    regex         = "([^@]+)@sha256:.*"
+    target_label  = "image"
+    replacement   = "$1"
+  }
+  // split image in name and tag
+  rule {
+    source_labels = ["image"]
+    regex         = "(.+):[^:/]+"
+    target_label  = "image_name"
+    replacement   = "$1"
+  }
+  rule {
+    source_labels = ["image"]
+    regex         = ".+:([^:/]+)"
+    target_label  = "image_tag"
+    replacement   = "$1"
+  }
 }
 
 prometheus.remote_write "prometheus" {
@@ -290,7 +315,35 @@ loki.source.docker "docker" {
 loki.source.journal "journal" {
   path = "/rootfs/var/log/journal"
   labels = { job = "{{ env "DOMAIN" }}" }
+  relabel_rules = loki.relabel.journal.rules
+  forward_to = [loki.process.journal.receiver]
+}
+
+loki.relabel "journal" {
+  forward_to = []
+
+  rule {
+    source_labels = ["__journal__systemd_unit"]
+    target_label  = "unit"
+  }
+  rule {
+    source_labels = ["__journal_syslog_identifier"]
+    target_label  = "ident"
+  }
+  rule {
+    source_labels = ["__journal_priority_keyword"]
+    target_label  = "priority"
+  }
+}
+
+// drop network db stats (15-20% of kernel logs)
+loki.process "journal" {
   forward_to = [loki.write.loki.receiver]
+
+  stage.drop {
+    expression          = ".*NetworkDB stats.*"
+    drop_counter_reason = "networkdb_stats"
+  }
 }
 {{ end }}
 

loki.yml.tmpl

@@ -15,7 +15,7 @@ ruler:
 
 server:
   http_listen_port: 3100
-  log_level: warn
+  log_level: {{ or (env "LOKI_LOG_LEVEL") "warn" }}
 
 distributor:
   ring:
@@ -56,12 +56,26 @@ schema_config:
       index:
         prefix: index_
         period: 24h
+{{ if env "LOKI_TSDB_FROM_DATE" }}
+    - from: {{ env "LOKI_TSDB_FROM_DATE" }}
+      store: tsdb
+      object_store: filesystem
+      schema: v13
+      index:
+        prefix: index_
+        period: 24h
+{{ end }}
 
 storage_config:
   boltdb_shipper:
     active_index_directory: /loki/boltdb-shipper-active
     cache_location: /loki/boltdb-shipper-cache
     cache_ttl: 24h         # Can be increased for faster performance over longer query periods, uses more disk space
+{{ if env "LOKI_TSDB_FROM_DATE" }}
+  tsdb_shipper:
+    active_index_directory: /loki/tsdb-index
+    cache_location: /loki/tsdb-cache
+{{ end }}
   filesystem:
     directory: /loki/chunks
 {{ end }}
@@ -74,6 +88,15 @@ schema_config:
     index:
       prefix: index_
       period: 24h
+{{ if env "LOKI_TSDB_FROM_DATE" }}
+  - from: {{ env "LOKI_TSDB_FROM_DATE" }}
+    store: tsdb
+    object_store: aws
+    schema: v13
+    index:
+      prefix: index_
+      period: 24h
+{{ end }}
 
 storage_config:
  boltdb_shipper:
@@ -81,6 +104,11 @@ storage_config:
    cache_location: /loki/index_cache
    resync_interval: 5s
    shared_store: aws
+{{ if env "LOKI_TSDB_FROM_DATE" }}
+ tsdb_shipper:
+   active_index_directory: /loki/tsdb-index
+   cache_location: /loki/tsdb-cache
+{{ end }}
 
  aws:
    endpoint: {{ env "LOKI_AWS_ENDPOINT" }}
@@ -99,7 +127,7 @@ storage_config:
 
 
 compactor:
-  working_directory: /loki/boltdb-shipper-compactor
+  working_directory: /loki/compactor
   compaction_interval: 10m
   retention_enabled: true
   retention_delete_delay: 2h
@@ -127,4 +155,4 @@ frontend:
 
 table_manager:
   retention_deletes_enabled: false
-  retention_period: 0s
+  retention_period: 0s

release/2.1.0+v1.18.1

@@ -0,0 +1 @@
+alloy: split image label in "image_name" and "image_tag", relabel journald logs and drop NetworkDB stats

release/2.2.0+v1.18.1

@@ -0,0 +1 @@
+allow scraping of docker events via new compose.docker-events.yml

release/2.3.0+v1.21.3

@@ -0,0 +1,2 @@
+- opt-in migration of Loki's index store from deprecated boltdb-shipper to TSDB via LOKI_TSDB_FROM_DATE
+- alloy and prometheus containers restart policy from on-failure to any

renovate.json

@@ -0,0 +1,6 @@
+{
+  "$schema": "https://docs.renovatebot.com/renovate-schema.json",
+  "extends": [
+    "config:recommended"
+  ]
+}