We are closing this incident due to system stability and will investigate for root cause.

The intervention at 18:40 UTC of rolling back to a stable version corresponded with system recovery that has remained stable since. We're investigating the root cause.

We rolled back to a previous stable version and we are monitoring.