Issue Summary

Jira DC currently has no size threshold for group deltas and does not emit an entity-level ENTITY_FULL_SYNC_NEEDED event for group membership changes.

Therefore, when Jira DC emits a normal group SAVE event containing approximately >10k membership removals,  the Connector Service sends the complete delta to JSWDD in one PATCH.

JSWDD fans it out into individual Group Directory membership DELETE operations with concurrency 5. The operation is subject to a 1,200-operation rate limit, so this delta is already much larger than the available quota.

As a result,  already-ingested Jira issues to be missing from Rovo search results.

Steps to Reproduce

1. Support has not been able to replicate this issue

Expected Results

Regardless of the data, users should be able to use Rovo and return issues

Actual Results

The below exception is thrown in the logs:

429 TOO_MANY_REQUESTS responses from the Unified Ingestion API’s group-membership SCIM PATCH path during jira-permissions-incremental-scan-task processing. 

Workaround

Currently there is no known workaround for this behavior. A workaround will be added here when available