split image tag

.env.sample

@@ -20,9 +20,6 @@ SECRET_BASIC_AUTH_VERSION=v1
 # server is remote
 # PROMETHEUS_REMOTE_WRITE_URL=https://prometheus.$DOMAIN/api/v1/write
 
-# Enable container filesystem usage metrics (expensive du-scan, ~30% CPU on large hosts)
-# CADVISOR_DISK_USAGE=1
-
 # Enable authenticated scraping of containers that opt in via
 # prometheus.io/auth=basic or prometheus.io/auth=bearer labels (used as
 # password/bearer token respectively). Insert it with:
@@ -87,7 +84,7 @@ SECRET_BASIC_AUTH_VERSION=v1
 # SECRET_GF_ADMINPASSWD_VERSION=v1
 ## Grafana's own domain. Defaults to $DOMAIN
 ## Change the value if you want Grafana on another domain.
-# GRAFANA_DOMAIN=$DOMAIN
+GRAFANA_DOMAIN=$DOMAIN
 #
 ## Single-Sign-On with OIDC
 # COMPOSE_FILE="$COMPOSE_FILE:compose.grafana-oidc.yml"
@@ -126,9 +123,3 @@ 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
-
-# Tell Traefik to keep access logs (could be very verbose)
-ALLOY_ACCESS_LOGS=false
-PROMETHEUS_ACCESS_LOGS=false
-LOKI_ACCESS_LOGS=false
-GRAFANA_ACCESS_LOGS=false

abra.sh

@@ -10,7 +10,7 @@ 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 CONFIG_ALLOY_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.

compose.alloy-webui.yml

@@ -14,4 +14,3 @@ services:
         - "traefik.http.routers.${STACK_NAME}-alloy.tls=true"
         - "traefik.http.routers.${STACK_NAME}-alloy.tls.certresolver=${LETS_ENCRYPT_ENV}"
         - "traefik.http.routers.${STACK_NAME}-alloy.middlewares=basicauth@file"
-        - "traefik.http.routers.${STACK_NAME}-alloy.observability.accesslogs=${ALLOY_ACCESS_LOGS:-false}"

compose.grafana.yml

@@ -43,7 +43,6 @@ services:
         - "traefik.http.routers.${STACK_NAME}-grafana.entrypoints=web-secure"
         - "traefik.http.routers.${STACK_NAME}-grafana.tls=true"
         - "traefik.http.routers.${STACK_NAME}-grafana.tls.certresolver=${LETS_ENCRYPT_ENV}"
-        - "traefik.http.routers.${STACK_NAME}-grafana.observability.accesslogs=${GRAFANA_ACCESS_LOGS:-false}"
     healthcheck:
       test: "wget -q http://localhost:3000/healthz -O/dev/null"
       interval: 5s

compose.loki.yml

@@ -34,7 +34,6 @@ 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"
-        - "traefik.http.routers.${STACK_NAME}-loki.observability.accesslogs=${LOKI_ACCESS_LOGS:-false}"
 
 
 configs:

compose.prometheus.yml

@@ -38,7 +38,6 @@ services:
         - "traefik.http.routers.${STACK_NAME}-prometheus.tls=true"
         - "traefik.http.routers.${STACK_NAME}-prometheus.tls.certresolver=${LETS_ENCRYPT_ENV}"
         - "traefik.http.routers.${STACK_NAME}-prometheus.middlewares=basicauth@file"
-        - "traefik.http.routers.${STACK_NAME}-prometheus.observability.accesslogs=${PROMETHEUS_ACCESS_LOGS:-false}"
 
 configs:
   prometheus_yml:

compose.yml

@@ -53,7 +53,7 @@ services:
         condition: on-failure
       labels:
         - "backupbot.backup=${ENABLE_BACKUPS:-true}"
-        - "coop-cloud.${STACK_NAME}.version=2.0.0+v1.18.1"
+        - "coop-cloud.${STACK_NAME}.version=1.6.0+v1.8.1"
 configs:
   config_alloy:
     template_driver: golang

config.alloy.tmpl

@@ -14,11 +14,7 @@ discovery.docker "linux" {
 {{ if ne (env "PROMETHEUS_REMOTE_WRITE_URL") "" }}
 prometheus.exporter.cadvisor "docker" {
   docker_only = true
-  {{ if eq (env "CADVISOR_DISK_USAGE") "1" }}
   enabled_metrics = ["cpu", "cpuLoad", "disk", "diskIO", "memory", "network"]
-  {{ else }}
-  enabled_metrics = ["cpu", "cpuLoad", "diskIO", "memory", "network"]
-  {{ end }}
   // host-wide totals already come from prometheus.exporter.unix
   disable_root_cgroup_stats = true
 }
@@ -52,7 +48,32 @@ 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" {

release/next

@@ -59,6 +59,3 @@ See the README's "Auto-discovering metrics from other apps" section.
 The Swarm, Stacks and Traefik dashboards were reworked to show old (pull-model)
 and new (Alloy push-model) data as one continuous line, so you don't lose history
 across the migration.
-
-The disk usage per container was disabled per default because of the CPU-expensive 
-filesystem scan (30% cpu increase). Enable by uncommenting env CADVISOR_DISK_USAGE=1