Hi! I believe I've found a bug where hitting the DB pool max connection limit and getting a legitimate KnexTimeoutError causes future queries to get KnexTimeoutError even when there are available connections in the pool. The latter, illegitimate KnexTimeoutError happens immediately when the query is made rather than after the timeout interval configured in the pool, which I think is telling that this is a bug somewhere.

It seems to happen in a very specific situation:

• Querying with withGraphJoined().
• Enough simultaneous queries to hit the DB pool max connection limit and have some queries time out while waiting for an available connection.
• Hitting the connection limit happens on the first query to the table, or something (I go into a little more detail about this in my comment in the repro case linked to below).

Here's my repro case: https://github.com/markdoliner/objection.js/blob/Repro_case_for_incorrect_timeout_bug/reproduction-template.js

I hoped to be able to track this down further but I don't think I'll be able to spend much time on it in the near future. My hunch is that there's somewhere in Objection where a promise from the first timed out query does not have appropriate error handling and is bleeding over into the subsequent query. But I could be wrong! Don't let me prevent you from exploring other possibilities.