Yesterday afternoon we experienced some problems with the service, initially problems with the search component of the service and shortly afterwards the site itself. The problems persisted for approximately 40 minutes. Unfortunately the initial SMS alert from the monitoring service either wasn’t sent or didn’t arrive, resulting in the on-call engineer only receiving the second alert half an hour after the first should have been sent, this meant the outage to the service was far longer than we’d normally expect. In order to try and avoid this kind of problem in the future, we’ve increased the frequency of monitoring alerts, so if one is missed, the subsequent message is sent sooner.