ResourceQuota Strategy

[kartoch]

Is your feature request related to a problem? Please describe.

A client has asked me to evaluate resource limits and requests at the namespace level. Currently, krr appears to only support evaluation at the deployment/statefulset/daemonset levels.

Describe the solution you'd like

A straightforward solution would be to calculate the resource limits and requests for each deployment/statefulset/daemonset and then multiply these by the maximum number of replicas. However, we must consider the rollout strategy. During a brief period, the number of pods could potentially double (in the case of the "Recreate" strategy) or increase by one (which is the default strategy).

Describe alternatives you've considered

I've thought about making the Prometheus requests manually.

Additional context

None, as far as I know.

[aantn]

@kartoch to handle the duplicates, could you first calculate the average usage and only then multiply by maximum replica count?

This isn't something we support now, but would you be interested in contributing a PR for it? Our team would be happy to assist and give pointers.