This incident was resolved.

Query error levels have returned to nominal levels, we continue to monitor.

We see another small increase in query errors.

Mitigations helped reduce number of query errors.

A small subset of query infrastructure nodes have crashed due to high load in one of our shards.