Loading commits from Fisheye inside a Jira issue times out after 10000ms
Platform notice: Server and Data Center only. This article only applies to Atlassian products on the Server and Data Center platforms.
Support for Server* products ended on February 15th 2024. If you are running a Server product, you can visit the Atlassian Server end of support announcement to review your migration options.
*Except Fisheye and Crucible
Problem
Fisheye commits fail to be displayed inside a Jira issue with the following message:
Unable to retrieve commit information
Jira is having difficulty contacting <FISHEYE_INSTANCE_NAME>. If this condition persists, please contact your Jira administrators.
The following message appears inside the <JIRA_HOME>/logs/atlassian-jira.log file when trying to load commits inside a Jira issue that fails with the message displayed above:
2020-08-17 15:55:47,735+0000 http-nio-8080-exec-138 WARN brosa 955x79561867x10 bcmmj6 127.0.0.1 /rest/dev-status/1.0/issue/detail [c.a.j.p.devstatus.provider.DefaultCoordinator] fisheye (timeout after 10000ms, not getting from cache)
2020-08-17 15:55:47,816+0000 http-nio-8080-exec-176 WARN brosa 955x79561869x11 bcmmj6 127.0.0.1 /rest/dev-status/1.0/issue/detail [c.a.j.p.devstatus.provider.DefaultCoordinator] fisheye (timeout after 10000ms, not getting from cache)Diagnosis
Environment
- Jira successfully integrated with Fisheye and Crucible using Application Links.
Diagnostic Steps
- You may see that Jira issues with a relatively small number of commits (< 10) don't exhibit any problems loading the commit details while Jira issues with a relatively large number of commits (> 150) do.
- While there's no restriction to the number of commits that can be linked to a Jira issue, there is a timeout in place for the Development panel of 10000ms (10 seconds). This is the time it has to retrieve all the information about commits linked to your Jira issue from Fisheye. If the operation takes longer than that, it times out.
The GET request made to <JIRA_BASE_URL>/rest/dev-status/1.0/issue/detail?issueId=3225186&applicationType=fecru&dataType=repository returns the following error when it times out (this can diagnosed by taking a HAR file when clicking the commits link under the Development panel)
error: timeout after 10000ms, not getting from cache
Cause
We are reaching the 10000ms timeout set for the Development panel when trying to load and display commits from Fisheye repositories associated to a Jira issue. The underlying reason why such timeout can occur may vary from deployment to deployment e.g. slow network, large number of commits associated to a given Jira issue and etc.
Workaround
It's important to note that this happens on the Jira side of things. The workaround must be applied to Jira, and not Fisheye and Crucible.
In order to workaround this problem we can increase the timeout set for the Development panel to give Jira a bit more time to retrieve the information about your commits. You can increase the timeout by adding the following JVM arguments to your Jira instance:
- -Djira.devstatus.coordinator.timeout.summary=30000
- -Djira.devstatus.coordinator.timeout.detail=30000
- -Djira.devstatus.capability.connection.timeout=30000
Please refer to the Setting properties and options on startup if you don't know where to add those JVM arguments.
Setting those arguments to 30000ms will increase the timeout from 10 seconds to 30 seconds. You'll need to restart Jira after adding those JVM arguments. If 30 seconds is not enough, you can increase the timeouts further if necessary.
In case the problem continues to happen even after increasing the timeout settings, please contact Atlassian Support.