Closed TiberiuGC closed 11 months ago
How about having a CLI flag called --resource-status
that accepts one of the following values[case insensitive] and filters the results accordingly? 💡
"CREATE_FAILED"
"DELETE_FAILED"
"UPDATE_FAILED"
"UPDATE_ROLLBACK_FAILED"
"ROLLBACK_FAILED"
How about having a CLI flag called
--resource-status
that accepts one of the following values[case insensitive] and filters the results accordingly? 💡"CREATE_FAILED" "DELETE_FAILED" "UPDATE_FAILED" "UPDATE_ROLLBACK_FAILED" "ROLLBACK_FAILED"
Yeah, I do like this one! 🚀
We should note that this new flag should be used with --events
flag. Not sure how useful it'll be on its own 😅
We should note that this new flag should be used with
--events
flag. Not sure how useful it'll be on its own 😅
Haha, we'll add some validations (and docs) for sure! 😄
Regarding integration tests - I've noticed that some of them use different cluster naming (e.g. trainium test) rather than the default params.ClusterName
. This results in the stack events not being printed when the suite fails, and instead we're seeing Error: unable to describe cluster control plane: operation error EKS: DescribeCluster, https response error StatusCode: 404
.
As part of this ticket, re-investigate all places in which we use params.LogStacksEventsOnFailure()
and ensure proper cluster naming is being set.
When running
eksctl utils describe-stacks
command, the user has the options to provide--events
flag in order to see a full event log as they would in the AWS Console. However, this could be considered noisy, depending on use case, and they may want to only filter for failed events. We could potentially provide this options via a new CLI flag e.g.--failed-only
and filter byResourceStatus
only for the following types:Note that we also run
eksctl utils describe-stacks --events
whenever an integration test spec fails, so adding this new flag there would probably be useful as-well.