Ben Sheldon [he/him]

fyi, I've removed the verification of the advisory lock (e.g. querying `pg_locks`) in the job performance: https://github.com/bensheldon/good_job/pull/1113 ...and released just now in `v3.20.0`

@ollym That query should no longer be taking place when jobs are dequeued/performed; it was removed entirely: https://github.com/bensheldon/good_job/pull/1113/files#diff-9d00af5d59d9a2b36d9ae265b052ec47618988f8d6e648b001901a7da8c3dc8aL493-R497 Are you still seeing the query being run after upgrading to v3.20.0?

I think what you're describing is the intended behavior, which is how I understand "Discarded" to mean: a job that has errored, but not been retried, is discarded. Active Job...

> if I pass a block [to `rescue_from`], I'm handling the error in the block and I want to treat it as rescued and handled. (I can always re-raise the...

Thanks for the back and forth. I guess just to really try to scope this down, is the core of what you're asking for: don't report an exception to the...

I spent a little more time tracing this through GoodJob. I'm surprised by this: > good_job.on_thread_error shouldn't be called. It looks like that _should_ be the case already if an...

@TAGraves that's really helpful context! (Also, I realized you also opened #844 which is related, though different) You are making me believe more strongly that I should record some context...

@TAGraves fyi, I started tracking the "error_event" in the database now (#995). I'm surfacing it slightly in the Dashboard, but there isn't currently any filters for drilling down into it....

@shouichi thank you for the feedback 💖 I'm hopeful that we can get to that level of granularity. Do you think the error_event types introduced in #995 cover all of...

It's weird that the behavior is different between `ActiveSupport::TestCase` and `ActionDispatch::IntegrationTest`. I think this is an issue caused by https://github.com/rails/rails/issues/37270 Could you try doing something like this in the `ActionDispatch::IntegrationTest`?...