OpenSearch for Jira early access program
The early access program (EAP) of OpenSearch for Jira gives you an opportunity to test apps' compatibility with OpenSearch.
If a plugin hasn't been migrated to the search API, it won't work with OpenSearch.
On this page:
Configure Jira to use OpenSearch
As Jira instances grow in size and scale, the default search engine, Lucene, may be slower to index and return search results. To address this, Jira Data Center now offers an alternative search engine as an opt-in feature—OpenSearch. OpenSearch can take advantage of multi-node instances to manage process-intensive indexing. In the user interface, the search experience remains the same for users.
Explore how to configure OpenSearch for Jira
Known limitations
Check out current limitations and known issues with OpenSearch support in Jira 11 Early Access Program (EAP). These limitations might affect your experience with searching, indexing, and overall performance:
- Sorting support is limited. Missing values might not be sorted as expected, and some sorting operations can fail.
- Unbounded queries aren’t supported. Check out the workaround for unbounded queries
- Read-after-write consistency isn’t enforced for all operations. Operations that delete documents won’t be consistent.
- Popular label suggestions aren’t available. Check out the workaround for popular label suggestions.
- Indexing isn’t resilient.
- Project re-index isn’t available. Check out the workaround for project re-index
- Only English indexing is supported.
- Indexing and re-indexing aren’t optimized for performance. Use small data sets.
- Background re-index isn’t supported for OpenSearch and will be removed in a future release. Use full re-index instead.
- The total results count is sometimes capped at 10,000.
To load Jira Software boards, increase max_result_window in OpenSearch. Check out the workaround for unbounded queries
Classes incorrectly annotated
In Jira 11.0.0, some new Search API classes were incorrectly annotated as PublicSpi.
These affected classes are implementations that aren't intended to be SPIs. They should have been annotated as PublicApi. Starting from Jira 11.0.1, all impacted classes listed below have been updated with the correct annotation.
com.atlassian.jira.search.Schemacom.atlassian.jira.search.ScoreComputedFieldSortcom.atlassian.jira.search.issue.index.indexers.impl.ExactTextCustomFieldIndexercom.atlassian.jira.search.issue.index.indexers.impl.LocalDateIndexercom.atlassian.jira.search.issue.index.indexers.impl.MultiUserCustomFieldIndexercom.atlassian.jira.search.issue.index.indexers.impl.NumberCustomFieldIndexercom.atlassian.jira.search.issue.index.indexers.impl.SortableTextCustomFieldIndexercom.atlassian.jira.search.issue.index.indexers.impl.UserCustomFieldIndexercom.atlassian.jira.search.issue.index.indexers.impl.VersionCustomFieldIndexe
Workarounds
Unbounded queries
By default, OpenSearch limits search results to 10,000 hits. Many queries in Jira attempt to retrieve all available results. These queries may fail until we replace them with more efficient alternatives, such as aggregation or streaming.
In the meantime, you can increase the OpenSearch maximum result window setting:
{
"index": {
"max_result_window": "2147483647"
}
}Popular labels suggestion
If you try to get the popular labels by selecting them from thje dropdown, you’ll get an error. Instead start typing the label you want to use.
Project re-index
Re-index the whole Jira instance instead of a single project.