fix: scrape once per hour

Following the discussion from the matrix channel, I made the disk usage metrics optional, as it currently causes ~35% of cpu usage on our system.

[Old cadvisor-config](02b01e5c23/compose.yml) from this repo had a `- "--housekeeping_interval=120s"` option enabled which seems currently not configurable for alloy - maybe this allows us reenabling disk metrics at a later point again.

dropped from 35% to 5% when removing `disk` from
`  enabled_metrics = ["cpu", "cpuLoad", "diskIO", "memory", "network"]` for cadvisor.

go profiling:

```
File: alloy
Build ID: de7ba8978f6753c25c13f6886ab3b896d3b05d3f
Type: cpu
Time: Sep 7, 2026 at 12:53pm (CEST)
Duration: 120s, Total samples = 37.23s (31.02%)
Showing nodes accounting for 32.20s, 86.49% of 37.23s total
Dropped 790 nodes (cum <= 0.19s)
      flat  flat%   sum%        cum   cum%
    28.72s 77.14% 77.14%     28.72s 77.14%  internal/runtime/syscall/linux.Syscall6
     0.47s  1.26% 78.40%      0.48s  1.29%  internal/filepathlite.(*lazybuf).append (inline)
     0.30s  0.81% 79.21%      0.94s  2.52%  internal/filepathlite.Clean
     0.26s   0.7% 79.91%      0.26s   0.7%  runtime.nextFreeFast (inline)
     0.23s  0.62% 80.53%      0.23s  0.62%  runtime.futex
     0.21s  0.56% 81.09%      0.21s  0.56%  runtime.memclrNoHeapPointers
     0.21s  0.56% 81.65%      0.21s  0.56%  runtime.memmove
     0.19s  0.51% 82.16%      6.19s 16.63%  os.(*File).readdir
     0.17s  0.46% 82.62%     31.24s 83.91%  path/filepath.walk
     0.12s  0.32% 82.94%      0.20s  0.54%  runtime.exitsyscall
     0.08s  0.21% 83.16%      0.24s  0.64%  runtime.scanObject
     0.08s  0.21% 83.37%      0.24s  0.64%  runtime.sweepone
     0.08s  0.21% 83.59%      0.29s  0.78%  slices.pdqsortOrdered[go.shape.string]
     0.07s  0.19% 83.78%      0.27s  0.73%  runtime.makeslicecopy
     0.06s  0.16% 83.94%     28.48s 76.50%  syscall.RawSyscall6
     0.05s  0.13% 84.07%      0.27s  0.73%  github.com/google/cadvisor/fs.GetDirUsage.func1
     0.05s  0.13% 84.21%     19.99s 53.69%  os.lstatNolog
     0.05s  0.13% 84.34%      1.26s  3.38%  runtime.mallocgc
     0.05s  0.13% 84.47%      0.43s  1.15%  runtime.mallocgcSmallNoscan
     0.04s  0.11% 84.58%      0.46s  1.24%  runtime.newobject
     0.04s  0.11% 84.69%      0.19s  0.51%  runtime.selectgo
     0.04s  0.11% 84.80%      1.18s  3.17%  runtime.systemstack
```

Reviewed-on: #29
Reviewed-by: Danny Groenewegen <247+dannygroenewegen@noreply.git.coopcloud.tech>
Co-authored-by: Simon <s.thiessen@local-it.org>

.env.sample

@@ -20,6 +20,9 @@ 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:
@@ -51,6 +54,9 @@ SECRET_BASIC_AUTH_VERSION=v1
 # SYSLOG=1
 # COMPOSE_FILE="$COMPOSE_FILE:compose.syslog.yml"
 
+# Monitor physical disks health
+# COMPOSE_FILE="$COMPOSE_FILE:compose.smartctl.yml"
+
 # Monitoring Server
 #
 ## Prometheus
@@ -84,7 +90,7 @@ 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"

README.md

@@ -166,3 +166,11 @@ 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`
+
+## 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

@@ -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=v1
+export CONFIG_ALLOY_VERSION=v2
 
 # 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.

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 +1,69 @@
+#! /bin/bash
+# Adapted from https://github.com/prometheus-community/smartctl_exporter/blob/master/collect-smartctl-json.sh
+
+script_dir=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
+
+# Data directory to dump smartctl output
+# This directory will be created if it doesn't exist
+data_dir="/var/lib/smartmontools/json"
+
+# The original script used --xall but that doesn't work
+# This matches the command in readSMARTctl()
+smartctl_args="--json --info --health --attributes --tolerance=verypermissive \
+--nocheck=standby --format=brief --log=error"
+
+# Ignore this devices
+smartctl_ignore_dev_regex="^(/dev/bus)"
+
+# Determine the json query tool to use
+if command -v jq >/dev/null; then
+	json_tool="jq"
+	json_args="--raw-output"
+elif command -v yq >/dev/null; then
+	json_tool="yq"
+	json_args="--unwrapScalar"
+else
+	echo -e "One of 'yq' or 'jq' is required. Please try again after \
+installing one of them"
+	exit 1
+fi
+
+if [[ ! "${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:"}
+	type="$(echo "${data}" | "${json_tool}" "${json_args}" '.device.type')"
+	family="$(echo "${data}" | "${json_tool}" "${json_args}" \
+'select(.model_family != null) | .model_family | sub(" |/" ; "_" ; "g")
+ | sub("\"|\\(|\\)" ; "" ; "g")')"
+	model="$(echo "${data}" | "${json_tool}" "${json_args}" \
+'.model_name | sub(" |/" ; "_" ; "g") | sub("\"|\\(|\\)" ; "" ; "g")')"
+	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.smartctl.yml

@@ -0,0 +1,18 @@
+---
+version: "3.8"
+services:
+  smartctl:
+    image: "prometheuscommunity/smartctl-exporter:v0.14.0"
+    volumes:
+    - "/dev:/dev"
+    - "/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

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

config.alloy.tmpl

@@ -14,7 +14,11 @@ 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
 }
@@ -48,32 +52,7 @@ 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/2.0.0+v1.18.1

@@ -59,3 +59,6 @@ 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