Attila Mészáros

So I see here more issues: 1. generating the RBAC 2. checking if required access rights are present on startup 3. integration with the probes.

> Problem with detecting access rights at startup is that it's equivalent to be able to generate the RBACs because you basically need to check that the operator has actually...

Added this functionality in for v4.1: https://javaoperatorsdk.io/docs/patterns-best-practices#stopping-or-not-operator-in-case-of-informer-errors-and-cache-sync-timeouts In #1594 a possibility is added to fine grain the liveness probe based on the health of the event sources / informers. These...

@metacosm can we close this issue now?

thx @shawkins for response! The hashing is over ResourceID because that is always at hand (note that these are not necessarily only Kubernetes resources), but can be external resources. And...

@lburgazzoli thanks for the feedback! I agree that using only the part 1 of the approach would need some careful measurement, and definitely does not sound that nice without the...

> I guess here we need to define what we mean by memory utilization because it is true that with sharding you can reduce the memory consumed by a single...

cc @salaboy (the discussion on our side regarding scaling mentioned on JBCN :) )

Checked it, now it actually creates a separate executor service, but this is more like a bug than a feature. When tested with this update that fixes that problem: https://github.com/java-operator-sdk/java-operator-sdk/pull/1371/commits/a2459dc432aee85768a2c4c6cfad1b167c1a7fa2...

Hi @andreaTP 1.the watched resource is accessible in multiple way, if you hold reference to the informer (variable), you can just pass read it using the resource id of the...