[kube-prometheus-stack] Add alertmanager persistentVolumeClaimRetentionPolicy #5078
Conversation
aslafy-z
changed the title
aslafy-z
changed the title
aslafy-z
requested review from
andrewgkew,
gianrubio,
gkarthiks,
GMartinez-Sisti,
jkroepke,
QuentinBisson,
scottrigby and
Xtigyro
as code owners
…onPolicy Signed-off-by: Zadkiel AHARONIAN <hello@zadkiel.fr> Signed-off-by: GitHub <noreply@github.com>
|
Could you hide this behind Capabilities.KubeVersion.Minor with version 1.32.0? This just got GA, I would rather not cause issues on existing deployments :) |
|
@QuentinBisson this just went GA but is enabled by default since its Beta phase, in 1.27. |
|
Oh yes you're right. Can you flag it for 1.27 to be safe ? I would assume we have people running out in later clusters |
|
@QuentinBisson I beleive this was already discussed at #4570 (comment). As the |
|
Oh my apologies, I was not aware of it :D Then yes, its good, let's merge it. Thank you for putting through with me today :P and for your work on this PR |
What this PR does / why we need it
This PR exposes persistentVolumeClaimRetentionPolicy field of Alertmanager CRD resource. PersistentVolumeClaimRetentionPolicy is a feature introduced for StatefulSets (by default in K8s 1.27.0+), this feature lets us decide whether the StatefulSet should delete or retain it's corresponding PVC's it is deleted or scaled. It is particularly useful in environments where stale volumes can grow in amount, causing additional costs or clutter in general. The CRD field itself was added to Prometheus operator with prometheus-operator/prometheus-operator#7183 and released with https://github.com/prometheus-operator/prometheus-operator/releases/tag/v0.79.0, however not yet exposed in kube-prometheus-stack.
Checklist
[prometheus-couchdb-exporter])