@toxinu Thanks. I wonder if your log parser had the same issue as mine? https://github.com/edx/configuration/pull/6089 I had to tweak it to get the log lines I showed above.
@nedbat I cannot confirm but it looks like it is a problem about when the logs are saved in your database vs when the actual email is sent (before the logs are saved in the database for instance). We have a fix being reviewed right now, so I will keep you in touch once it is fixed.
@nedbat, as @toxinu mentioned we have a fix which is waiting for review.
Since we would like to provide as much useful information as we can, in an easily readable way, I’d like to ask you to check the “Structure of the email” section of the Pull Request. I did change it a bit to be more “efficient”.
In case you have any suggestions related to the email or its format/structure, feel free to comment that on the PR.
FTR: The fix is deployed and I can see the new emails coming in.
The CI is working well to alert us to issues. Can we get it running on the open-release/koa.master
branches also?
@nedbat, yes, we can. I’ll reserve time in our upcoming sprint to set this up for open-release/koa.master
.
Any progress on the Koa CI? Is their something we can do to help get it launched?
@nedbat, finally we updated our instance manager and now it works nice with Ubuntu 20.04 (ref), so CI is ready now.
When provisioning is finished, you’ll be able to find a fresh Koa instance using this address: https://periodic-build-koa.opencraft.hosting.
It will be pulling and deploying changes from open-release/koa.master
every 12 hours.
Am I understanding right that failure messages should be sent to open-edx-btr-notifications@googlegroups.com address? If so, could you check the inbox? You should see first message.
Yes, I am getting messages about Koa, but they say they are failing:
{'changed': False, 'msg': "No package matching 'ack-grep' is available"}
BTW, for quick scanning in the inbox, it would be cool if we could adjust the subject line. Now it is either:
Deployment back to normal at instance: Periodic build Ironwood (periodic-builds-ironwood.opencraft.hosting)
Deployment failed at instance: Periodic build Ironwood (periodic-builds-ironwood.opencraft.hosting)
Better would be to shorten the subject and move the important information closer to the front:
Deploy Ironwood CI: Passed
Deploy Ironwood CI: Failed
No need for “back to normal”, especially since if it always passes it will always say back to normal.
Yes, a few builds failed with this error, but then the issue disappeared. It’s related to BTR-40. We’re not sure what can be the reason. Has anyone else faced this?
Sure, I’ll plan this for the next week. Thanks for the feedback
@nedbat, is it fine if subject will be slightly different? Like:
Periodic build Koa CI: Passed (or Failed)
By the way, here is the pull request:
@nedbat FYI this PR was merged and deployed, so you should have started receiving emails with updated subject. Could you confirm?
Yup: “Periodic build Juniper CI: Passed”. Thanks!