This issue has been resolved, you can see the overview for further details around what occurred during the incident.
Overview
- The Raygun customer web application was unavailable to a portion of our customers for 4 hours 23 minutes on 9 June 2021 between 2:22pm and 6:45pm UTC.
- It was caused by high load on a single application server, and a fault in communication between that application server and our primary database.
- Our alerting and paging systems worked as planned, but the error was incorrectly believed to have been remedied when CPU load dropped to normal levels.
- Our load balancer did not take the application server out of rotation as the health check did not ascertain whether the database was able to be accessed from that specific application server.
- No crash reporting, RUM or APM data was lost during this period.
- Some deployment registrations and sourcemap/symbol uploads may need to be retried.