Crowdcast experienced an issue where hosts appeared live and recording, but attendees could not view the broadcast. Engineers quickly identified the cause as an overloaded database. To restore service, they switched database configurations, resulting in a brief (~5 minute) site interruption before the event was reset and service recovered.

We are improving infrastructure provisioning for large events to prevent similar issues in the future.