Intermittent request failures on EU cluster

Incident Report for Pusher

Resolved

Between the hours of 02:30 and 09:30 this morning a node in our EU cluster became degraded and automated health-checking failed to remove it from service correctly.

This resulted in a proportion of client requests to both the HTTP API and the WS API to receive authentication errors for valid requests to the EU cluster.

Only customers of the EU cluster were affected.

We are investigating the failure of the health-checkers to route around the failed node and apologise for any inconvenience caused.

Posted Jan 11, 2016 - 10:26 UTC

This incident affected: Channels REST API and Channels WebSocket client API.