Nils Bandener

@scrawfor99 For increasing visibility, do you think whether it might make sense to tag this issue with the `performance` label?

@peternied > As you mention there are a diverse set of cluster configurations that would work better with different implementations to get around this what do you think about adding...

@anijain-Amazon Thanks for your comments! > Can we pre-compute this data asynchronously as well based on changes in user roles and requested objects ? > Similar to how it's being...

> Will there be an eviction logic for all the data structures mentioned in the document ? The data structure mentioned in the document so far, are not intended to...

> On the data structure side, would it be necessary to store an entry for each concrete index if there are many indices that match a pattern and permit the...

So, here are some test results for heap usage of the `Map` data structure. I used this test program: ``` import java.util.HashMap; import java.util.HashSet; import java.util.Map; import java.util.Set; import org.openjdk.jol.info.GraphLayout;...

@DarshitChanpura > This would be easy for a fresh cluster, however, how do we plan to approach this for existing clusters. For the basic (non-DLS/FLS) privilege evaluation, the logic is...

> There won't be any performance issues as they sit in memory and there will be no I/O retrieval/storage involved. The updates will be similar to ConfigUpdateAction and the listener...

@DarshitChanpura > Do you think we are in a good place to get a POC with an in-memory data structure you proposed along with LRU implementation? Absolutely. One open thing...

@peternied I have added my thoughts at https://github.com/opensearch-project/security/issues/3903#issuecomment-1938822478