The 'outages' column is showing all times where our monitoring service wasn't able to reach the named service. This includes numerous releases I performed to the main website. The who list and the main website are technically the same thing. The Jira system can be slow to respond and you see that reflected as it being down more often than it really is. And Sindo.Me was purposefully left offline early this month as part of a test of the monitoring system.
Overall, I am pleased with the performance and stability of our systems. I am hoping that when we move to the new server, the new process management solution I'm building will reduce the length of time it takes to restart key processes like the main website application. This should end up reducing the number of 'outages' that get detected as I make code changes.
