So what are all these 9s about. You always hear about this five 9s or two 9s, They’re usually talking about the uptime that you have. And it really resolves into how much unplanned or even planned downtime that you can have over a period of time. So it’s basically the risk exposure that you adopt by adding an extra 9 goes down by an order of magnitude each time. Now you do it yourself, how do you keep your cost from going up by an order of magnitude. It’s all about architecture and planning and understanding how to do this stuff. So an SLA, talking to a technical team of people here. We all know what APIs are as programmers. A SLA is an API for your business. It’s how you interact with the people you’re doing business with. It’s when I’m paying you, what are you guaranteeing me. And you’ve all seen the Microsoft end user license agreement, that’s a SLA. On the eJamming service that I was telling you about that has not had any downtime, just a couple of weeks ago, their drive filled up, and we got a call. The things broken again, first time we looked to see is, sure enough they filled the driver. They were suppose to fix this the last time, the time before but they haven’t done it. And when your file share fills up, there’s only one thing to do. You must reset your system, because the Linux kernel depends, it has certain assumptions about availability and accessibility for your drive share and it just gets knocked crazy. You’ll never going to have a safe recovery on that, so you actually have to reboot the machine. And if you’re really cool geek ops guy, what you really pride yourself on is what your uptime says. The uptime of that server was 436 days. And that’s because last time it filled up. Literally that time of the server should’ve been like 3, 4, 5 years. That’s when you know you have a system that’s stable. You’re never going to do this on a Windows environment.
At time 0:38, the speaker means ‘down’ instead of ‘up’