Description

Currently operator deployments have a memory limit of 128Mi.

On very large clusters with hundreds of namespaces this is not enough and leads to the operator Pods getting OOMKilled.

Patching the operator memory limits is especially tedious on OLM deployments because the UI doesn't have any mechanism to customize this prior to the installation.

Acceptance

• All operator Helm charts have been updated to the new limits.
• TODO: decide on the exact new value for those. 256Mi seems to be good for ~500 namespaces per cluster