Downtime Postmortem - 11:43-11:45AM (ET)

Incident Report for Greenhouse

Resolved

Our hosting provider Heroku accidentally triggered a database restart on our systems at 11:43AM ET earlier today. This caused a period of unavailability lasting approximately two minutes as our systems automatically cut over to our backup database. No data was lost or changed as a result of the failover to our backup. This restart was triggered without any coordination between Heroku and Greenhouse staff--a state of affairs which is obviously unacceptable for both us and for our customers.

We are working closely with Heroku's senior leadership team to ensure this sort of thing does not happen again. Longer-term, we will be moving our database services out of Heroku and onto AWS where we will maintain more control. We apologize for any issues this brief downtime may have precipitated, and deeply appreciate your understanding and your business.

Posted Jan 27, 2015 - 18:25 UTC