You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This commit was created on GitHub.com and signed with GitHub’s verified signature.
* test: remove flaky basic_expiry integration test
* test: fix flaky test_route_with_one_healthy_and_one_unhealthy_seed
Replace fixed sleep durations with retry loops that actively poll for expected routing state. This eliminates timing-related flakiness in CI environments where task scheduling is less predictable.
Changes:
- Add wait_for_routing_to_domains helper that polls until expected domains are available
- Update test to use retry loops with timeouts instead of fixed sleeps
- Remove timing assumptions that failed in CI due to scheduler variance
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* test: improve wait_for_routing_to_domains reliability
Increase sample sizes and timeouts to account for CI environment variability:
- Use 10 samples for single domain, 30 for multiple (matching other tests)
- Calculate timeouts based on snapshot_update_duration (fetch + 2*check intervals)
- Increase poll interval to 100ms to reduce overhead
These changes align with the timing patterns used in other routing tests and provide sufficient margin for CI scheduler variance.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
* Revert "test: improve wait_for_routing_to_domains reliability"
Revert "test: fix flaky test_route_with_one_healthy_and_one_unhealthy_seed"
These commits attempted to fix the flaky test with retry logic, but the test continues to fail in CI. Reverting to consider alternative approaches.
---------
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>