This incident has been resolved.

Service has been fully recovered and we're continuing to monitor.

We've identified and reverted the change that is causing the issue, and are in the process of recovering the systems affected. Existing projects are unaffected.