You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Right now the-new-hotness adds a lot of comments (which in turn generates a lot of emails) when it files an issue. It would be nicer to combine everything into the initial report, or perhaps the initial report and a single comment.
The downside to pushing everything into a single report is that there's a large period of time where something can go wrong/the service is restarted and we lose the state and fail to report the new version.
The text was updated successfully, but these errors were encountered:
1 comment if anything happens during start of the scratch build
1 comment for attachment being added
1 comment for output of the scratch build
We can easily put together 1 and 2, we just add the bugzilla comment after scratch build was started (with the current implementation, we always get exception when something happens during the start of scratch build). 3 is currently not possible to merge with the previous one, because it's separate bugzilla API call. And the 4 is happening asynchronously to previous one.
So the best we can do right now is put 1 and 2 together. Which will lower the amount of updates by one. Not ideal, but something to go with.
With Redis implementation of the cache we can probably save the whole notification till the scratch build finishes and then do just two updates - one for the whole comment and one for attachment. Then the maximum amount of updates will be 2 even with scratch build enabled, but the cons here is that the notification will be available later than now, after scratch build will be finished.
Right now the-new-hotness adds a lot of comments (which in turn generates a lot of emails) when it files an issue. It would be nicer to combine everything into the initial report, or perhaps the initial report and a single comment.
The downside to pushing everything into a single report is that there's a large period of time where something can go wrong/the service is restarted and we lose the state and fail to report the new version.
The text was updated successfully, but these errors were encountered: