It can be concluded that one of our databases suffered from a degradation in the performance of the underlying storage layer. An automatic failover to a hot standby was automatically performed. The failover happened after a period with very high write latency. We are in contact with our service provider to determine why the failover did not happen faster. The are looking into it, but they have assured us that an incident like this is extremely rare.
The incident affected the ability to create new payments for alternative payment methods (non-card) for approximately 10 minutes.
We apologise for the inconvenience.