On March 14th, 2019 01:37 UTC, our monitoring system detected an outage on our Accounts application. The service was impacted for a total of 29 mins. During this time users were not able to access their account or billing details and were unable to provision new projects.
Investigation revealed a cache that failed to properly clear was responsible for the issue. A restart of the database resolved this, restoring the service at 02:06 UTC.
As a preemptive action, we are improving our cache clearing configuration to ensure we do not encounter the same issue in the future.