hieradata/common/prometheus.yaml

Loading that file relies on the wmflib::expand_path backend which is not in use in Cloud VPS.

I'm ok to move prometheus::instances_default elsewhere in hieradata where it would be compatible with cloudvps (not sure where?) Alternatively could we add hieradata/common to cloudvps (or beta's puppetserver) hiera ?

Change #1144589 had a related patch set uploaded (by Elukey; author: Elukey):

[operations/puppet@production] Add prometheus::instance_defaults to deployment-prep's common settings

https://gerrit.wikimedia.org/r/1144589

The patch would set:

prometheus::instances_defaults:
  retention_time: 4032h
  retention_size: ~
  thanos_upload: true
  k8s_cluster_name: ~
  hosts: ~
  provision_lv_size: '50g'

But then puppet error becomes did not find a value for the name 'prometheus::instances', so there is some mismatch still.

The prometheus::instances: settings seem to also be in common/prometheus.yaml and apparently describe various Prometheus clusters.

Mentioned in SAL (#wikimedia-releng) [2025-05-12T15:21:59Z] <bd808> Added prometheus::instances and prometheus::instances_defaults hiera settings to "deployment-etcd" Prefix Puppet via Horizon (T393866)

In T393866#10812265, @Stashbot wrote:

Mentioned in SAL (#wikimedia-releng) [2025-05-12T15:21:59Z] <bd808> Added prometheus::instances and prometheus::instances_defaults hiera settings to "deployment-etcd" Prefix Puppet via Horizon (T393866)

Applied config:

prometheus::instances: {}
prometheus::instances_defaults:
  hosts: null
  k8s_cluster_name: null
  provision_lv_size: 50g
  retention_size: null
  retention_time: 4032h
  thanos_upload: true

Mentioned in SAL (#wikimedia-releng) [2025-05-12T15:28:03Z] <bd808> Forced puppet run on deployment-etcd02.deployment-prep.eqiad1.wikimedia.cloud to fix Puppet run (T393866)

Mentioned in SAL (#wikimedia-releng) [2025-05-12T15:28:08Z] <bd808> Forced puppet run on deployment-etcd05.deployment-prep.eqiad1.wikimedia.cloud to fix Puppet run (T393866)

Change #1144589 abandoned by Elukey:

[operations/puppet@production] Add prometheus::instance_defaults to cloud's common settings

https://gerrit.wikimedia.org/r/1144589