Closed tomas-zijdemans-vipps closed 11 months ago
I believe we may have introduced a bug in function dispatching as we ramp up the feature rollout to address bumping the timeout for locally-running functions to 15 seconds, to match the deployed timeout (#179). It seems the logic in our solution may have inadvertently affected deployed functions when calculating the timeout. We are reverting this change as we have seen timeouts and retries increase across our entire system.
I will post an update once the revert is complete and I have an update from our side / our monitoring.
We have rolled back the change and the number of timeout (and thus requeues/retries) has dropped dramatically.
I think that was the underlying cause of the issue.
I will close this issue but if you see it happen again please re-open and we can continue the investigation.
Thanks!
Hi,
So something really weird happened this morning (08:40 Oslo time). A workflow suddenly behaved very weird, was there a deploy in the backend? I've seen this sort of issue happen before randomly.
It seems like the following is happening:
My workflow code containing 4 steps:
Slack logs: