Preventing wasted resources

Hi all, to prevent lots of wasted resources by using the istio-sidecars, the best way forward is to use custom annotations? I can use these annotations on any resource? (namespaces, deployments, pods) and they will be applied from bottom (pod) to top (namespace) ? or what’s the best approach to finetune resource consumption?

How can I calculate the optimal CPU/Memory for the istio proxies?

Am I correct to say that the istio sidecar CPU/Memory annotations only apply on POD level? This means we need to configure them on ALL workload pods (and the ones deployed through Helm charts) So meaning: updating all deployments + Helm charts ?

Thanks :slight_smile: