An error in our logging system caused the instance buffer to be filled and hence was causing the instances to crash. We restarted the instances and cleared the buffer to bring the system back. In parallel, we fixed the issue with the logging system.

We are observing outage of multiple core APIs and investigating the root cause.