Slicehost is smarter than I am
So this afternoon I took a little break from running unit tests to head to the store and pick up a couple things. When I got back, I noticed I was no longer on IRC, and my client was reporting it couldn’t connect to the bouncer I run to stay online. I tried to SSH in to see if the bouncer had died, and discovered I couldn’t connect.
And couldn’t ping the box.
And couldn’t get this site to come up in a browser.
This site’s hosted at SliceHost on a nice beefy VPS, and they provide a nice web-based interface for checking on server status, so I logged in there and found I couldn’t get a console to come up. Diagnostics reported that the server was indeed down and said support had already been contacted.
At which point I popped open my inbox and saw not one but two emails from Slicehost support: one telling me their monitoring had noticed the server was down, and another a few minutes later saying there’d been some sort of hardware problem and they were swapping it out.
I was, it goes without saying, impressed; I’m not used to consumer hosting noticing and working on an issue before the consumer (in this case, me) is even aware of the problem, but that’s what they’d apparently been doing.
So I sat tight for a bit, and then saw another email come in telling me that the hardware swap was done, and the server was being brought back online. A minute later I had SSH access again, and watched the box do the usual confused dance of bringing all its services back up; a couple minutes later, everything was running and stable again.
Elapsed time: about thirty minutes. For nearly fifteen of them I wasn’t even aware of any problems, but the Slicehost folks were and were on it. That’s what I call service.