From 3ea485bfc79d9a4e3218ad6714c1c68ea0eff9f4 Mon Sep 17 00:00:00 2001
From: Simon <s.thiessen@local-it.org>
Date: Tue, 8 Sep 2026 10:30:59 +0200
Subject: [PATCH] make disk usage per container metric optional

---
 .env.sample       | 5 ++++-
 abra.sh           | 2 +-
 config.alloy.tmpl | 4 ++++
 release/next      | 3 +++
 4 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/.env.sample b/.env.sample
index 40f1612..c70c433 100644
--- a/.env.sample
+++ b/.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:
@@ -84,7 +87,7 @@ 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"
diff --git a/abra.sh b/abra.sh
index 2f27db8..f9f90b2 100644
--- a/abra.sh
+++ b/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.
diff --git a/config.alloy.tmpl b/config.alloy.tmpl
index c8af4ad..f775a53 100644
--- a/config.alloy.tmpl
+++ b/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
 }
diff --git a/release/next b/release/next
index 7226256..e7c348f 100644
--- a/release/next
+++ b/release/next
@@ -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
-- 
2.54.0