Performance degraded below our performance standards between 11:50 and 13:50 PDT today. Response time in this period averaged approximately 1200ms.

Following the rebrand this morning, we renamed and reconfigured many services to clean up old usages of the "Beam" branding. The systemd service which ran our backend API was configured to run a smaller number of Node.js processes than necessary to fully utilize our application servers. This caused response times to slow, even though server load averages and other metrics remained acceptably low. The incident was immediately resolved after raising the number of worker processes.
May 25, 14:09 PDT