Avoid “Resource update cancelled” as UPDATE_FAILED reason

When a resource update fails, concurrently updated resources also fail
with “Resource update cancelled” reason, and these events happen after
the actual culprit. Since events are scanned backwards, skip such
“cancelled” failures, increasing the likelihood of reporting the actual
stack update failure reason.

Author: artyom

main.go

@@ -126,7 +126,7 @@ func run(ctx context.Context, stackName string, args []string) error {
 				if evt.ClientRequestToken == nil || *evt.ClientRequestToken != token {
 					continue
 				}
-				if evt.ResourceStatus == types.ResourceStatusUpdateFailed {
+				if evt.ResourceStatus == types.ResourceStatusUpdateFailed && aws.ToString(evt.ResourceStatusReason) != "Resource update cancelled" {
 					return fmt.Errorf("%v: %s", evt.ResourceStatus, aws.ToString(evt.ResourceStatusReason))
 				}
 				debugf("%s\t%s\t%v", aws.ToString(evt.ResourceType), aws.ToString(evt.LogicalResourceId), evt.ResourceStatus)