e2e: Increase all ANR timeouts to 2m to ensure CI reliability. #1733
Conversation
| // start is async, so wait some time for cluster health | ||
| time.Sleep(time.Minute) |
There was a problem hiding this comment.
Was this sleep just never needed? Like a minute long sleep is long
There was a problem hiding this comment.
Maybe it was needed in the past, but it doesn't seem to be needed now.
There was a problem hiding this comment.
And to be clear, this sleep ensured each e2e job was wasting 30-40s. At least with a timeout the check can complete as soon as the nodes are ready, but this sleep will never exit early if the nodes are healthy earlier than expected.
There was a problem hiding this comment.
The missing piece here is understanding that Health actually blocks until Healthy.
There was a problem hiding this comment.
yup that was very old ANR tech dep. with the first grpc server implementation, start (with blockchain creation -I know this is not the case but this is a copy probably-) broked posterior health call without sleeping some time.
StephenButtolph
added
ci
Why this should be merged
CI seems to be exceeding many ANR-related timeouts. Rather than bumping timeouts piecemeal, all ANR timeouts are set to the same constant of 2 minutes.
How this works
How this was tested