Oban seems to be bringing down my application about oban HOT 6 CLOSED

After this was reported I asked @axelson a few questions on Slack and diagnosed the issue. There are cascading failures do to how the notifier manages its connection, and how the producers rely on establishing listeners on start up.

I have a set of sweeping fixes that I'm working on and aim to have them merged today.

from oban.

The fix turned out to be rather involved!

I think the resulting architecture is much cleaner though, processes are more isolated and far more resilient. Locally I can stop postgres entirely, watch the circuit breaker messages scroll by for a few minutes, then turn it back on and everything returns to normal after ~30 seconds.

from oban.

Thanks for the update!

from oban.

@sorentwo Is it possible to get this in a new release soon? (and thanks for the quick fix!)

from oban.

@axelson Absolutely. v0.10.0 is published now.

from oban.

Great, thanks!

from oban.