Cloud Intel currently relies on the tenantId being prefixed to the db name in the slow query log (), so we manually ensure that it's always prefixed. Eventually, they will change their application such that it does not rely on this - once they do, we should stop logging the tenantId as the prefix in this log line.