feat: garage metrics coop-cloud/garage#8 #22

2026-06-11T22:20:36Z

fauno commented

2026-06-11 22:20:36 +00:00

since metrics by default are protected by a token, they can't be exported locally like we're discussing at #20

fauno added 1 commit 2026-06-11 22:20:37 +00:00

feat: garage metrics coop-cloud/garage#8 fd5ed0a1d3

fauno added 1 commit 2026-06-11 22:34:06 +00:00

fix: enable secret 1561b97c54

fauno added 1 commit 2026-06-11 22:38:40 +00:00

fix: shorter secret name toolshed/abra#867 7f9a1beed1

p4u1 commented

2026-06-12 06:50:47 +00:00

How would I scrape multiple garage instances?

p4u1 commented

2026-06-12 08:22:05 +00:00

https://grafana.com/docs/alloy/latest/reference/components/discovery/discovery.docker/

It seems it is possible to configure a bearer token for the discovery with docker labels

https://grafana.com/docs/alloy/latest/reference/components/discovery/discovery.docker/ It seems it is possible to configure a bearer token for the discovery with docker labels

dannygroenewegen commented

2026-06-14 20:04:21 +00:00

I was thinking about a label-based approach, see the alloy-updates branch commit e247677433. Discovering the targets that have to be scraped dynamically based on container labels. I haven't tested this yet.

Any container can opt in to being scraped by adding Docker labels:

labels:
  - "prometheus.io/scrape=true"
  - "prometheus.io/port=3903"       # Garage metrics port
  - "prometheus.io/path=/metrics"   # optional, /metrics is the default
  - "prometheus.io/auth=basic"      # optional, uses the basic_auth secret

Alloy then discovers all containers, filters to those with prometheus.io/scrape=true, and scrapes them automatically. I've only added noauth and basic auth, so for Garage, prometheus.io/auth=bearer should probably be added. Multiple Garage instances on the same host would each be picked up as separate targets by their service name. Alloy just needs to be configured with all needed types of auth and other possible parameters, and then the container labels do the rest. No need to add every app we want to scrape into monitoring-ng.

I was thinking about a label-based approach, see the alloy-updates branch commit e247677433. Discovering the targets that have to be scraped dynamically based on container labels. I haven't tested this yet. Any container can opt in to being scraped by adding Docker labels: ``` labels: - "prometheus.io/scrape=true" - "prometheus.io/port=3903" # Garage metrics port - "prometheus.io/path=/metrics" # optional, /metrics is the default - "prometheus.io/auth=basic" # optional, uses the basic_auth secret ``` Alloy then discovers all containers, filters to those with prometheus.io/scrape=true, and scrapes them automatically. I've only added noauth and basic auth, so for Garage, `prometheus.io/auth=bearer` should probably be added. Multiple Garage instances on the same host would each be picked up as separate targets by their service name. Alloy just needs to be configured with all needed types of auth and other possible parameters, and then the container labels do the rest. No need to add every app we want to scrape into monitoring-ng.

❤️ 1

fauno commented

2026-06-20 16:37:01 +00:00

already implemented metrics discovery

fauno closed this pull request

2026-06-20 16:37:01 +00:00

Pull request closed

Please reopen this pull request to perform a merge.

Sign in to join this conversation.

3 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: coop-cloud/monitoring-ng#22