On 6 January, as usage of Teem services increased for the day, there was an increase of duplication errors on our database. This increase caused a delay in replication, which in turn caused a greater number of duplication errors. This negative feed forward mechanism caused a degraded site performance. Reduced load on the database by scaling down our service instances allowed the replication to catch up, and remain health. Going forward, there will be database maintenance on the specific tables that exhibit these issues to guard against this in the future.