Issue Summary

In some Jira environments, the Instance Optimizer Index Auto-Healing Feature can run for an excessive duration without actually completing any indexing tasks.

Most of the time is spent analyzing the index, while attempts to repair it are repeatedly unsuccessful. The built-in fail-safe is not triggered, causing the process to churn for hours through 100 iterations.

Steps to Reproduce

1. Jira DC instance with index auto-healing enabled and a large issue count (>1.0 million issues).
2. Introduce an orphaned issue in jiraissue (only set id and updated).
3. Set com.atlassian.jira.indexhealer to DEBUG.
4. Run the scheduled job, or repeatedly trigger manually via POST /rest/indexhealer/1/heal

Expected Results

After 3 iterations of the Instance Optimizer's indexing auto-healing feature with no reduction in the unhealthy count, the fail-safe trips, and the job exits early.

Actual Results

Every iteration logs Successfully healed: N. Failed to heal: 0. Remaining 1 to heal.

The job runs all 100 batches and ends with Total items failed: 0  alongside Total items unhealed: 1.

Workaround

Schedule the Integrity Checker to run checks on a shorter schedule to prevent the index healer from becoming caught in a loop.