Issue 138128 failure

[benchaplin]

This PR addresses the flakiness of SearchWithRandomDisconnectsIT.testSearchWithRandomDisconnects (closes elastic#138128), which was observed to time out or hang on CI.

The test had two main issues:

1. Timeout: The disruption loop, which runs 50-100 iterations each followed by ensureFullyConnectedCluster(), lacked a time bound. On slower CI environments, this could cause the test to exceed the 20-minute suite timeout.
2. Hang: During cluster teardown (often triggered by the suite timeout), runMoreSearches() could call prepareRandomSearch(), which would then throw a RuntimeException("Cluster is already closed"). This uncaught exception prevented finishFuture from completing, leading to future.get() blocking indefinitely.

The fix implements the following changes:

• Added a try-catch block around prepareRandomSearch().execute(f) within runMoreSearches() to ensure that finishFuture is always completed, even if prepareRandomSearch() throws an exception (e.g., due to a closed cluster). This prevents the test from hanging.
• Introduced a 5-minute time limit for the disruption loop to ensure it completes within a reasonable timeframe and does not contribute to exceeding the overall suite timeout.
• Unmuted the test in muted-tests.yml.

---

  

[cursor]

Cursor Agent can help with this pull request. Just @cursor in comments and I'll start working on changes in this branch.
_{Learn more about Cursor Agents}