This test takes ~25s when run in the aggregation suite on AL2 with all feature flags, and ~2m30s when run in the implicit_query_settings_fallback suite.

Under TSAN (no feature flags), it takes 10m30s in aggregation and ~1hr in implicit_query_settings_fallback. With all feature flags, the runtime pushes even higher to the point where it times out.

We should figure out a mitigation, increase resource limits, or decide to denylist the test.