Michael J. Sullivan

Ahhhhh, it looks like `session_idle_transaction_timeout` is always included in `DESCRIBE INSTANCE CONFIG` because during bootstrap, we execute CONFIGUREs for any backend settings with default values: https://github.com/edgedb/edgedb/blob/df39a451f8c49efc2a332fb4f4072d7b20b9807c/edb/server/bootstrap.py#L1761-L1782 We don't currently have...

That's kind of wild. @fantix any idea what could cause this?

I think part of the reason for the difference with postgres is that we are using postgres's SERIALIZABLE isolation level, which is stricter than the default of READ COMMITTED. When...

Hmmmm, I had thought that `TransactionSerializationError` was marked as `SHOULD_RETRY` also, but it is not: https://github.com/edgedb/edgedb/blob/b9b4177d8c0080235e46b7e825384224fbe5923c/edb/api/errors.txt#L128C18-L128C47 Maybe we need to change that Doing synchronous locking should work great if you...

Oh, good point. It looks like the rust bindings *do* reflect that faithfully: https://github.com/edgedb/edgedb-rust/blob/093729876737738a559e35c2ec3f0ae27efd2f69/edgedb-errors/src/kinds.rs#L139. This sort of counter workload is not a great fit for SERIALIZED/REPEATABLE READ, perhaps.

EdgeDB sets the mode in the connection settings when it connects to postgres, I think. Changing it to `read committed` there does fix this problem, though I don't think there...

Here's one where it isn't getting cleared right: https://github.com/edgedb/edgedb/actions/runs/9025625652/job/24801671577?pr=7327

I think it is a race with the scheduled reintrospection

What should we do to test this? Some tests that talk to the language server directly?

Getting rid of the branch semaphore takes me from 13.5m for the tests (not counting a minute for cached bootstrap) to 11.5.