Make the dashboard more efficient

### Context

I have been using the dashboard a lot recently while debugging the failed processor jobs.

### Problem or idea

The dashboard still takes like a minute to refresh the cache and times out.

### Solution or next step

Are all of our queries actually necessary? Are we hitting the index properly? Would implementing https://github.com/AlexsLemonade/refinebio/issues/1842 help?

Start by tracing the dashboard cache refresh that currently takes about a minute and times out, then inspect whether its queries use the expected indexes. Review issue #1842 as a possible related direction. Done means the dashboard refresh completes without timing out and the required queries and index usage are understood.

Written by the indexing model from the issue text.