> Update - primary failover briefly improved performance but did not fully mitigate, we've throttled inbound traffic and are investigating upstream Vitess issues