Not good at all, we apologize for the incident!
Approximately 10:20 CET today 2015-06-03 one of our backend database services stopped responding. This affected the majority of our external services. We where degraded/partial outage for about 20 minutes.
The reason was that our provider was doing rolling upgrade of the database machines and during rolling restarts one host machines load went up too high and they had to stop it.
So the upgrade was planned but the downtime was not.
This was a wake up call for us that we put so much trust in this service. We will work actively at removing this single point of failure.