Intermittent API timeouts

Incident Report for Sanity

Resolved

The intermittent API timeouts were caused by a node crash due to a bug in the Linux kernel's CFS scheduler. A kernel upgrade has been scheduled and will be carried out within the next week. This affected roughly 1.5% of requests within the given time period.

Posted Sep 29, 2017 - 09:27 UTC

Monitoring

Users may have experienced intermittent API timeouts from about 08:42 to 08:48 UTC, possibly due to network issues. We are monitoring the situation and investigating the underlying cause.

Posted Sep 29, 2017 - 08:57 UTC