Jira occasionally times out during execution of workflow post function

[Chris Cain]

My organization has developed a Jira plugin (running in 4.3.3) that provides workflow post functions to call a remote server to execute code there. In a small number of cases Jira is timing out during execution of this post function before receiving a response from the remote script. The error message displayed is as follows:

"The call to the JIRA server did not complete within the timeout period. We are unsure of the result of this operation.

Close this dialog and press refresh in your browser."

This seems straightforward enough - except that we have scanned the Jira source code and been unable to locate the source of this error message or how to configure the timeout period. Any assistance with this would be much appreciated.

[Jobin Kuruvilla [Adaptavist] · Accepted solution]

Looks like there is a timeout configured for ajax xalls in atlassian-jira/includes/js/ajs/smartajax.js file. Check out the function at the bottom. It specifies a timeout 60000ms.

Try increasing it and see if it helps but I must warn you it is there for a reason. You should probably try to find out why the call is taking more than a minute!

Make sure you modify the minnified version of the js as well.

[Chris Cain]

Thanks for the advice, I'll coordinate this modification with the admin of the Jira server in question and see how it goes.

[Brad Baker [Atlassian]]

I have to agree that more than 1 minute is an awful long time to be taking to update an issue. Sensible update times would be in the 100-1500 ms range and the top end of that is on the high side.

Can I ask what is taking that long to complete? Maybe there is a better way to do that.

[Chris Cain]

The reason that the workflow transition takes so long in some cases is that we are using Jira to front end a system that manages code deployments to (potentially) multiple platforms. While there are definitely targets for improving efficiency, the bigger the code changes involved, the longer the workflow transition will take. If this architecture proves unsuitable I may investigate the possibility of making this asynchronous - but we don't want the Jira issue status to transition until we know whether or not the code deployment completed successfully.

[JamieA]

Maybe consider creating a quartz job for immediate execution in your workflow function. If the deployment fails you could update the issue with a comment, or revert the status. Either of those will send a mail so the user knows there is a problem.