Hi there,
I’m testing my Jira Server to Cloud migration processing and it goes through fine. However, the progress never goes to 100% and the status stays “Running” indefinitely as you see on the image below, in that case the migration was finished already couple of minutes ago and progress notification was sent to the cloud endpoint:
I do send the progress to the /rest/atlassian-connect/1/migration/progress/0e7707a7-9215-349c-9791-2977319baaeb - endpoint and my last post there when the migration processing is finished consists of a body with
{
progressStatus: ‘SUCCESS’,
statusMessage: ‘All diagrams fixed!’,
percent: 100
}
I would assume that it should complete the migration progress, but it doesn’t seem to do that from the users point of view. This is a problem in that the user never gets the 100% completion of the migration process, even though everything has already been migrated properly.
Is this a known bug with JCMA (version 1.5.4) or is there something else I should be doing to make the migration progress 100% and completed?
Regards,
I have the same issue with the Jenkins Integration for Jira app.I think it could be because the server app and the testing cloud app have different app keys and there is an open bug that the JCMA app doesn’t correctly handle this. In some cases the cloud app key is ignored and the server app key is assumes to match the cloud app key.
I saw in testing that the progress was only shown if the server and cloud app keys where the same. The issue was logged by Atlassian, but the I only kept the link in theSlack workspace that has now been closed.
I’ll try to locate it to share with you.
Cheers,
Mark
@markrekveld
Well, in my case both keys are the same so shouldn’t be this issue:
private static final String SERVER_APP_KEY = "com.gliffy.integration.jira";
private static final String CLOUD_APP_KEY = "com.gliffy.integration.jira";
Cheers,
That makes it even more interesting. Seems that the closing the Slack workspace may have been a bit too soon.
I did find the bug MIG-592
Hi @PetriJuhaniRiipinen,
I can see the status update requests for your transfer ID is processed successfully from our logs.
Status updates from the cloud apps are not going to be reflected in CCMA or JCMA real time. I’ll discuss this with the team and come back with updates on this.
Could you please confirm whether the app migration status gets updated to the expected state after some delay (say 10 mins) ?
Hi @HariharanRajendran,
Well, 10min was a good guess… I just ran a new test transfer (id = 4121f5e8-814a-3876-9e7c-ca93e0992102) and once my code posted the 100% SUCCESS-message, I started a timer and waited for 10 min. Just when the 10min was fulfilled, the Progress changed to 100% and the green checkmark appeared there.
I think I posted some intermediate progress as well between 0% and 100% but those are then lost as only 0% and then 100% after 10min are going to be shown to the user. Sounds to me that this progress is a bit useless feature for any intermediate progress indication if the user isn’t going to see any of that anyway.
Regards,
1 Like
Hi @PetriJuhaniRiipinen,
Thanks for confirming the behaviour. I’ll take this to team and come back to you with any plans about improving app migration status update experience.
Thanks and regards,
Hariharan.
1 Like
Hi @PetriJuhaniRiipinen ,
Status updates feature is not designed to be real time and there can be up-to 10 mins delay between actual status update from a cloud app and corresponding status update reflection in CMAs UI.
As each app can have different interval between their status updates, our decision is to start with a longer interval for status update reflection, observe status update intervals
of different applications and change the status update reflection interval to an optimal value based on our observation.
We understand this delay in status update reflection can add some overhead in testing your application’s status update feature. We are discussing about updating the status update reflection interval to a shorter duration when running in dev mode. We’ll post an update regarding this next week.
2 Likes
Hi @HariharanRajendran
Ok, I do see your point there. But what about this: When the progress of SUCCESS with 100% (or even just SUCCESS) is sent, maybe then update the migration progress on GCONF as quickly as possible as that should be the final progress update during the application migration.
It’s not really an issue for my local testing (although enabling it to work faster with dev mode sounds good) as I know what’s happening and I do see my own debug logging which tells me when the migration is finished. I’m more thinking of the customers admin persons who are running the migrations, they may be confused when the progress seems to be 0% for undetermined period of time and then jump directly to 100% 10min after the migration actually finishes. That may cause unnecessary support requests either to you or us, which could be avoided by improving the progress indication speed.
Regards,
2 Likes
Hi @HariharanRajendran,
I observed one more issue, when calling post request for progress api, getting 403 error frequently. Even if we are trying 3 times we are getting 403 errors. We are calling using axios post.
Any suggestion on this.
2 Likes
Yep, exactly same observation here, lots of 403 errors with Axios post to progress API.
Hi @DharmaTeja and @PetriJuhaniRiipinen,
If you are getting 403, it also means that transfer may have already been settled.
Settled means either SUCCESS/FAILED/INCOMPLETE status has been sent previously, thus, the further updates wouldn’t be allowed.
Please ensure that your service isn’t making subsequent settled request, after the first one.
If above isn’t the case, then please let us know when that happens, and will check the logs to find out more about the issue.
Thank you,
Anshuman
1 Like
