This morning at 7:45 AM PT there was a 5 minute slow down on job processing caused by some reverse proxies being overloaded due to their autoscaling being disabled as part of maintenance. During this time running tasks and loading sandboxes may have performed more slowly than usual.
The issue quickly resolved itself and we've taken steps to prevent future occurrences. Specifically, we've more than doubled the minimum number of these proxies in use and completed the maintenance, hence re-enabling auto-scaling.
Posted 7 months ago. Feb 13, 2019 - 07:45 PST
This incident affected: Project API and Sandbox Cluster.