Persistent Volumes Questions

[vapomalov]

Can somebody elaborate why there is now (v2.3.0) a use for persistent volumes for registry, chartmuseum, jobservice, redis and trivy?
It seems because of the size they are used for caching?
We are hosting harbor in aws eks, so without readwritemany we will need to switch to efs which is way slower than ebs. Will the performance of efs be sufficient for the usecase of the persistent volumes?
Or does somebody know an alternative solution where pods are not bound to one availability zone?