Closed saumier closed 6 months ago
@troughc @sahalali This is the issue blocking events coming from Co-motion to Footlight CMS that Maude complained about this morning.
The issue related to the aggregator was the latest firewall update done in the production server. That issue is fixed and all the 4 actions ran successfully.
I am now trying to see, how to make the "action failed" while any unexpected error occurred while running the script. So that we will be informed when fail.
@sahalali I have setup a test in the Ville de Laval workflow with an invalid CMS API ( -i https://apissss.cms.footlight.io). I will check on Thursday if the workflow failed and if I received an email.
If yes, then I will close this issue.
BTW, thanks for breaking long lines into multiple lines. Much much appreciated and easier to read and edit.
test worked. I'm closing.
None of the aggregators have run successfully since Nov 2nd. When you look at the time to complete all of the recent workflow jobs over the past 3 weeks, the time to run drops substantially to less than a minute, whereas before Nov 2 is was taking several minutes.
Please investigate. An idea that pops into mind is the Nginx firewall for Admin API.
Expected behaviour: If there is a 403 Forbidden error during the job run, then the entire job should fail. Github has a rich set of features to notify users when jobs fail. Other types of errors should also cause the job to fail, but not a 408 "Request Timeout" during a batch call to Artsdata when there can sometimes be an individual batch failure due to a timeout.
Current behaviour: One of the steps in the job has an Axios 403 Forbidden error and the job still passes and we are not aware of a problem.