Ankit Jain

> There is great work in search performance - and views impacts reducing overhead for privilege evaluation. When a search request is created, such as `GET indexes-*/_search` the permissions evaluation...

> So a user could run a query GET my-alias/_search could return 200 and then an admin changes the underlying pointer and it starts to return 403 - not because...

> In my mind, there is an existing conceptual model that are users are aware of, GET BLANK/_search it works in a very specific way and if we change that...

> So a user could run a query GET my-alias/_search could return 200 and then an admin changes the underlying pointer and it starts to return 403 - not because...

> If the view is updated to point to a different index (or indices), then, yes, the user would be able to query that different index (through the view). I...

> @msfroh @reta @jainankitk This is really good discussion. I've created an RFC [1] to discuss the problem space - I think that will lead to better alignment before I...

@ruai0511 - Thank you for creating this PR. Can you update the title/description with recent name (query group) and schema.

@ansjcy - Thank you describing the detailed solution as part of this PR. As per my understanding, the customer/client should not even be supplying the label. Instead based on some...

> @jainankitk -- can you please run through your comments and resolve the ones that have been addressed or highlight the ones that still need work? Thanks! @ansjcy - It...

> Did extensive benchmark tests, merged the tests results by calculating average on multiple runs, and here are the test results: I don't see significant impact on search latency with...