New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Reconsider Horizontal Pod Autoscaler docs for CephObjectStore #10001
Comments
I am not quite sure how reconciliation will impact HPA, by default HPA can work with memory and CPU with any special configuration, so I highly doubt reconciliation of resources can affect HPA.
During CI testing it validates all the yamls in |
From my rook-operator restart testing, Rook operator resets the replica count 1 as you mentioned and I guess HPA increases it suddenly to previous value replica value for the workload
|
We should update our CephObjectStore reconciliation. I think a small design would make sense. Questions I have:
|
@thotz KEDA updates the replica count on the rgw deployment, right? So all Rook will really notice during a reconcile is that the deployment replica count is incorrect? In that case, I'm thinking we should do the following to keep it simple to support KEDA:
Another consideration is a new property to the object CR for |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed in a week if no further activity occurs. Thank you for your contributions. |
Not a wontfix, but probably not able to prioritize until 1.11+ |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed in a week if no further activity occurs. Thank you for your contributions. |
This issue has been automatically closed due to inactivity. Please re-open if this still requires investigation. |
I am concerned the documentation for KEDA Horizontal Pod Autoscaler (HPA) documented in the CephObjectStore won't work in production. I believe any time the CephObjectStore is reconciled, Rook will reset the replicas to 1. This could result in disrupting user workloads.
Originally posted by @BlaineEXE in #9202 (review)
We should investigate whether this is a concern and reasses. I believe the example is a good one, but we may need to design a Rook feature to assist with this use-case. If we decide not to proceed with a Rook feature and there is an issue, we may want to revert the commits.
In any scenario, we should remove the KEDA CRDs from the current location in
github-action-helper.sh
since they are of no use to our CI.The text was updated successfully, but these errors were encountered: