We are seeing severe delay in events delivery to forge triggers (on-delete, on-assigned mostly) - exceeding ~40 minutes for each event. Forge docs states about no more than 3 minutes.
Has anyone observing this? Is there some kind of investigation we can do to troubleshoot this? Or the platform is dead?
Such delays render whole forge trigger mechanism useless.
Probably due to this incident that is now resolved: Atlassian Developer Status - Delay in Jira Webhook Delivery
Yeah, this was it.
However there is another case to bear in mind. When forge trigger function does await for its promises - next event deliveries (thus calls the trigger function) may by disrupted. I thought that maybe there is another similar case we are not aware of.
We already reported this, although for Confluence: https://jira.atlassian.com/browse/ECO-1403
It is still very much an issue for the whole platform, not just during that incident.
Whenever a shard gets overloaded, customers can and will receive triggers severely delayed.