Background job processing went down on 10/1 or us-east-1 region and 10/4 for eu-west-1 region. Both incidents were due to the same root cause: * A key scheduling service failed. * This was triggered by our core transactional database recycling after a minor version upgrade. * This happened around 3am eastern time on 10/4 in us-east-1 region. * Happened around 3am eastern time on 10/1 in eu-west-1 region. * We should have noticed this, however there was an error in some of our alerting / notification scripts. * We have since updated and thoroughly tested the notification scripts.