All times are shown in UTC
We are seeing high error rates and latencies worldwide, due to issues with our Cassandra persistence layer instances in ap-southeast-1 and us-east-1. We are investigating.
Latencies and error rates are now back to normal. The issue was routine maintenance in our Cassandra persistence layer that caused significantly more work than expected, leading to contention issues. We will consider using a different strategy moving forwards to reduce the amount of work needed during maintenance.
Resolvedabout 2 months ago