Index issues from Jira Cloud projects, with their comments. Each issue becomes one document, and citations link back to it.
This is for Jira software and work projects. For service desk requests, see Jira Service Management.

​
Before you start

  • A Pro plan or higher, and permission to edit sources in the knowledge base.
  • A Jira Cloud site, for example https://acme.atlassian.net.
  • An Atlassian account that can browse the projects you want indexed. Kelu sees exactly what that account sees, so a dedicated account is a good idea.

​
Set up

1

Create an API token

Sign in as the account Kelu should use and create a token at id.atlassian.com.
2

Add the source

In Kelu, open your knowledge base, go to Sources, click Add source and choose Jira. Give it a name, then enter the Base URL, the account’s Email and the API token. To index some projects only, list their Project keys.
3

Test and add

Click Test connection. It checks that the account can see at least one project. Then click Add source. The first sync starts right away.
An issue tracker rarely belongs on a public site. Documents from a new source are public by default. To keep them to your team, add Jira to an internal knowledge base, or put it in its own source group and keep public surfaces away from it.

​
Options

The add form shows the connection and Project keys. The filters are on the source’s Configuration tab. Each empty filter keeps everything.
OptionDefaultWhat it does
Project keysEmptyComma-separated, for example ENG, PLATFORM. Empty indexes every project the account can see.
StatusesEmptyOne per line, spelled as your workflow spells them, for example Done.
Issue typesEmptyOne per line, for example Bug or Story.
ResolutionsEmptyOne per line, for example Fixed. Keeps only issues resolved that way, which keeps out Won't Do and Duplicate.
Created afterEmptyA date as YYYY-MM-DD. Only issues created on or after it.
Updated withinEmptyOnly issues changed in the last N days.
Only issues that reached a resolutionOffKeeps any resolved issue. Ignored when Resolutions is filled in.
Filters combine: an issue must match every filter you set.

​
What gets indexed

  • One document per issue, titled KEY-123: summary.
  • It holds the type, status, resolution, priority, labels, components, fix versions and dates, then the description and the comments.
  • Rich text keeps its code blocks, links and tables.
  • Long threads keep the newest 200 comments. The source’s Sync history tab lists any issue that was cut.
  • Attachments are not indexed.
When you narrow a filter, the next sync removes issues that no longer match. Updated within is different: an issue that ages out of the window stays indexed.

​
Sync

Jira re-syncs every 10 minutes. Issues that did not change are skipped. Click Sync Now on the source page to sync right away.

​
Troubleshooting

  • The sync fails with status 401 or 403. The token was revoked, or the account lost access. Create a new token and update it on the Configuration tab.
  • the credential works but reaches no project. The account can sign in but browse no project. Give it access to the projects.
  • Fewer issues than expected. A project key or filter value is misspelled, or the account cannot see that project.
  • The sync fails and says to narrow the source. One sync reads up to 500 pages of 100 issues. Add project keys, statuses, issue types or a date filter.
  • created_after "…" is not a YYYY-MM-DD date. Fix the Created after value, for example 2024-01-01.