At 6.00am PST on 16 October Workspace had an outage. The first customer report was received at roughly 7.30am PST and the Workspace bat-phone received its first call. The server was restarted which brought Workspace back online as of 9.20am PST.
Enabled the http profiler to provide details that enabled us to fully investigate the 'goroutine leak' in the workspace backend that we were seeing on production servers only.
Improve http client connection handling to eliminate the resource leak that caused the Workspace outage on 16 October.