Performance issues in the North American region resulted in connection timeouts to containers in our infrastructure, which caused some partitions to be in a hidden failed state.

The affected instances were identified and restarted to restore normal operations.