"It's not just making good things for users or making good things for the company you work for, but making good things that don't wake up your coworkers in the middle of the night."
DevOps for the Masses
DevOps for the Masses
Chris Corriere is a DevOps Engineer at Autotrader, a mathematician, scientist and a columnist at DevOps.com. He is also one of the co-organizers of DevOpsDays Atlanta, responsible for growing the DevOps community in the SouthEastern United States. His work at AutoTrader puts him into a position to build a dashboard for a new way to look at metrics and measurements for evaluating the efficiency of DevOps initiatives.
In today's journey, we talk with Chris about his early influences, his community work with DevOpsDays Atlanta and what he hopes to leave as a legacy to the community.
Mark Miller: When did you first get started with technology?
Chris Corriere: I was really young. I can remember playing games on an IBM PC Junior when I was in pre-school. I moved into the technician phase before I was in middle school. I was taking things apart and putting them back together.
Mark Miller: Did you have a whole group of people around you, or were you flying solo?
Chris Corriere: A lot of it was just driven out of my own curiosity. I did have an uncle that was a double E and had a background in mathematics. I had gotten into circuits and some of the Forrest Mims books out of RadioShack and building different little TTL things and making LEDs turn on. He was glad to have somebody to talk shop with. I spent a good bit of my childhood trying to figure out some of the stuff he was telling me at the time. He was definitely a large inspiration for me.
Mark Miller: Where did you start getting your education as far as technology is concerned? Did you go to school after that?
Chris Corriere: A lot of what I learned was self-taught and reading online. I started coding back when GeoCities was around and had my first website there. Eventually I moved up to Atlanta to go to school at Southern Polytechnic State University, which is now part of Kennesaw State. I learned a lot at Southern Poly. I came in there as a double E. I moved into computer science for a little bit before I finally ended up in the math department.
I also got hired by the Division of Information Technology to do a lot of lab support and engineering programs. I was working with a lot of high-end CAD machines and neat lab equipment like potentiostats and gas chromotographers. There were machines hooked into wind tunnels. I got a lot of unique experience working in that environment.
Mark Miller: When did you first recognize DevOps? Do you remember the first time you heard it?
Chris Corriere: There's really three different things I think of when that comes up.
I was in high school, I think I was in my junior year. My dad pulled me out of classes and sent me to a workshop they had in my hometown, Albany, Georgia. The Economic Development Committee had brought in Amazon to do a talk on e-commerce. I had gone in to that expecting to come out knowing how to build an e-commerce website, a deep dive into the technology. They spent the entire time talking about supply chain management, about getting books delivered and figuring out what your inventory is.
At the time, I remember being immediately disappointed. Seeing the other people in the room perk up and start to pay more attention, I started thinking this is probably relevant. I should probably take notes. That's been a steady pattern I've seen throughout my career in this space is focus on quality, you got to look at supply chain.
From there, when I got on campus at Southern Poly ... There is a brief period of time when people really were able to not believe in antivirus and did not run it on most personal computers. There was a summer when Lovebug and klez, a few others came out. Then the virus problem really got bad and consumers really had to have antivirus on their systems. They had to be more conscious of security. During that time we had some people in the lab that were able to go and delete out viruses by deleting registry keys in windows. When it mutated or a different version of it came out, we were back to running antivirus scans, which is time-consuming. Even when you get them cleaned up the machines don't always work quite exactly the same way they did before.
It was about that time we took disk-imaging practices and created base images and a process to back the machine up and reload it and migrate the data back from the backup on to the machine in about an hour. At that point, we were able to time box our lab and desktop support. If you were going to spend more than an hour on something, back it up and re-image it before you did anything else. I didn't know about MTTR (Mean Time to Repair) then. We weren't tracking it. But I did see a backlog of 70 to 100 tickets drop down to three in less than six months. We did run in to issues from time to time where that didn't solve everything and there was something else going on.
That was where I really first ran in to the concept of disposable infrastructure. We were really maintaining applications as a service and the data that the users had around those. But the individual computer that they were on on campus and that was really pretty arbitrary. If they moved quick, they liked it.
Mark Miller: What year was that?
Chris Corriere: That would have been 2003, 2004.
Mark Miller: So that's pre DevOps visibility. You were doing it, but it didn't have a name yet?
Chris Corriere: Well and we were doing this with desktop support on a college campus. It wasn't in cloud-based infrastructure. That sort of leads to the next step.
I had moved in to consulting from there and did that in Atlanta for a couple of years. Went and worked for a .NET shop and wrote some C# for them and then went to Home Depot and was supporting their grid. They had a Tomcat grid.
It was at that point, I had a manager hand me a copy of The Phoenix Project. I remember him telling me, "I can't quite tell you what DevOps is, but you definitely do this thing. You should read this book." I was pretty much instantly hooked after I got through The Phoenix Project.
Mark Miller: You're attributing your career now to Gene Kim?
Chris Corriere: Yeah. Without hesitation. There's direct, irrefutable evidence.
Mark Miller: What's your motivation now for working in DevOps? You've been at it for years now. What's the motivation to keep going?
Chris Corriere: Gene Kim, again. It's improving the lives of ten thousand IT professionals I think is the line. The longer I do this the more whittled down that statement gets and the broader the base it applies to. I think this is improving lives. I think these are practices that help us leverage technology in the best way possible and to be more helpful.
I think that dips in to the Pareto Efficient Nash Equilibrium where it's not just making good things for users or making good things for the company you work for, but making good things that don't wake up your coworkers in the middle of the night. That work-life balance gets pulled into it as well.
Mark Miller: When you're thinking back over how you got started, what do people need nowadays if they want to move in to the DevOps track?
Chris Corriere: You've got to move towards experimentation. Get your hands in to code and that's source code or infrastructure's code. It should probably be some of each. Start in the shallow-end of the pool, get something simple set up, run a proof of concept on a deployment pipeline. Get some perspective and experience firsthand. I think where a lot of people trip up is trying to figure out what tool or what language or what platform they should set something up on. Try one. If it's out there, people are using it and you learn something.
Mark Miller: It's interesting as I talk to people about the whole DevOps movement. One of the things that keeps coming up over and over is the idea that in order to be a good DevOps practitioner you have to be able to be a good communicator. You had to be able to work and show teams how to collaborate. Do you buy into that?
Chris Corriere: I do believe communication in key. That dips back in to Conway's law right? That how we communicate is inevitably going to effect whatever it is we're producing. That the systems we build are going to model our communication structure.
The more I work with and learn about micro services the more I realize it has little to nothing to do with the API so much. Those are a piece of the technology that is a vehicle of communication. It's the bus we use for that. Micro services is really a team structure and it's that communication and the trust that goes along with that communication that's critical to the success, more so than the platform you're building it on.
Mark Miller: When you look back on your accomplishments, what are you most proud of so far?
Chris Corriere: That's a tough question. I'm going to have to say DevOps Days Atlanta. We just wrapped that up, that's a pretty immediate impression in my memory still. That's been on my list of things to do, things that I was hoping to see in Atlanta one way or another.
We had the event, that was my index where it was the check mark. Did it happen and it did. We sold it out. We got a lot good feedback, lot of good talks, really good time. Looks like we're set for next year too. That's exciting. I'm looking forward to doing it again and try to help do some more community building in the Atlanta area.
Mark Miller: Of the projects that you've worked on, say with your recent gig with AutoTrader, what's going on there? Is there something that you're really happy that you're working on there?
Chris Corriere: I've been putting together a dashboard framework that was really more a stopgap for me to get me out of manual reporting that I inherited. We've been able to automate a lot of metrics with it. I think that's about to hit next level where pulling in data into context and helping users make educated decisions with real-time up-to-date information.
Most of the time when I start talking about this people assume it's Graphite or ELK Stack or Splunk. There's a lot more metrics involved and we're measuring things in a lot of different places that I'm not sure they're doing other places yet. I'm interested to get some more feedback on that but it's been positive so far. Seeing how that's intersecting with our automated testing framework and how that's evolved since I've started, I think things are headed in a positive direction.
Mark Miller: I have to say that John Willis speaks very highly of you. How have you been working with John?
Chris Corriere: We met at Velocity a few years ago and talked there some. I've seen him at a few meetups in Atlanta. We actually don't live too far from each other so we catch up for lunch from time to time.
We have a lot of similar concerns. He's been talking a lot about burnout. That's something I've noticed in he industry as well I think is a problem. We need to do a better job of addressing. It goes back in to the cultural aspects of this and the work/life balance. We've both got kids. I'm a big advocate, I believe everybody can learn how to get into technology. I've come to accept the fact that not everybody wants to and that's fine. I do a lot of work with my kids helping them to learn to code and really looking towards the future of this sector and what technology's going to look like in ten years. Making sure we're doing the best we can with what we've got.
A big thanks to all the people who are in this community and help make it awesome, John Willis and Gene Kim for sure. There's too many people to name. Bridget Kromhout and Jennifer Davis do a lot of awesome stuff. I've enjoyed working with them over the past year. I'm glad to be a part of this community.
Mark Miller: What legacy would you like to leave with the community? When you look back on your career, what would you like to be able to say you've accomplished?
Chris Corriere: That is hard to do. I'd like to be known for helping people understand each other better and to understand themselves better I think is the first step to that. It's sort of chicken before the egg… how well do you understand yourself before you start talking to other people about that.
We do what we can and learn from our mistakes and keep plugging along forward. I get a lot of happiness out of that, seeing people get together, seeing people meet and have a space where they can talk about technology with people who've lived it and dealt with some of these successes and failures firsthand. I think it's important that we have a welcoming space to do that. People don't have a doubt that we'll be glad to see them if they show up.
Mark Miller: If you could make up your own title to tell people what you do, what would the title be?
Chris Corriere: It's hard to say. I like DevOps Engineer because it's pretty broad-based. It's adaptable. I am very focused on automation in and around data-driven decisions. I wouldn't take offence to be called a data scientist. But I think scientist works well. I think that accurately describes me.
I'm a mathematician at heart but I don't think that resonates with most people. I'm an applied mathematician so we can go with scientist.