Replies: 1 comment
-
There is an issue for this that if implemented would solve so many other issues than with HPA. See fluxcd/pkg#696 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Scenario:
There is a known behavior when Deployments with HPA are used with Flux where Flux reconciliation overwrites the HPA replica settings in deployment. The workaround suggested by HPA and Flux in this FAQ is to remove the .spec.replicas from the Deployment spec.
Issue:
For existing deployments in production, removing
.spec.replicas
results in a temporary degradation where number of replicas is reduced to one. The recommendation in HPA docs to workaround this degradation is to transfer ownership when using server-side apply. While the suggested workaround is possible when there are manual cluster operators with a few clusters, it is not feasible with a large fleet of clusters with thousands of deployments deployed using Flux.Question:
Is there an alternate way to address this when HPA is used with deployments? Is it feasible to detect the conflict and remove the replica configuration when the controller applies the diff ?
Beta Was this translation helpful? Give feedback.
All reactions