If your primary role within your organization is Operations, you likely realize the importance and significance of monitoring. The DevOps philosophy teaches us that measuring and collecting as much as possible is critical to not only the health of infrastructure, but it also plays a critical role in the architecting efforts of both software and hardware.
Jason Dixon of Librato knows this better than many and has made a career of helping others understand and implement the best tools and processes of successful monitoring. Known by many and first introduced to me as the “Graphite Guy”, Jason’s expertise goes far beyond his contributions to this and other open source projects.
Monitorama, a popular tech conference organized the last 2 years by Jason, focuses on the best practices and tools used by experts in this field and has been hugely popular among those in the Web Operations community. Jason and I first got to know each other last year during the 2014 Velocity conference in New York City. Since then, he’s been a valuable resource in our ongoing partnership with ingesting Librato monitoring data to alert the right people or teams via VictorOps.
Here’s a bit more about Jason and what he’s been up to as well as what is planned for the remainder of the year…
Q: For those out there who are unfamiliar with who you are, can you give a quick intro?
Sure, my name is Jason Dixon and I lead the Integrations team at Librato. Aside from my day work, some folks know me from my monitoring (and DevOps-related) open source projects (Tasseo, Graphite, and Synthesize, to name a few).
I’m the founder of the Monitorama conference series and I also happen to be writing a book about Graphite for O’Reilly. Technically, I reside in Maryland but my heart lives on the west coast; Portland and San Francisco, to be precise.
Q: Where do you feel DevOps (the big picture) is now… compared to this time last year?
I think people are finally (years late, imho) coming to the realization that “DevOps? What’s that?” doesn’t matter so much as getting real work done in a collaborative manner. To be fair, I don’t know that I’ve ever fully bought into the community side of DevOps so much as I wished that people would just focus on enlightening the conversation and breaking down walls between all internal teams, not just Operations and Engineering.
In many ways though, I’m probably not the target audience; I love tools and seeing how they can be improved to enhance our workflows, rather than sitting around talking about processes for hours on end. I have a hard time saying that without a fair bit of snark, but I honestly believe we’ve spent far too long talking about “What is DevOps?” instead of areas that really need improving, such as diversity in tech, and more specifically, the ridiculous disparity between men and women in Operations.
Q: Any tools/services that you’ve heard a little about .. and want to learn more?
Well, I think the “proper” answer here is to talk about the hot containerization technologies like Docker, Rocket, and Mesos. Ironically (as a tools fan), it seems to me we’re way too focused on the tooling here without taking a good, hard look at the user experience.
It’s been four years since I left my job at Heroku (an eternity in internet time), and yet many of our developers are using the same shitty tools today that they have been for years. Projects like Cloud Foundry and OpenStack have been trying to push Enterprise PaaS down everyone’s throats without (seemingly) a solid understanding of what makes the Heroku experience so appealing.
I bet there’s still a ton of engineers out there who have no idea that Heroku uses LXC for the application slugs (runtimes), only that they can create and deploy a new app, add a database, and scale it up instantly, in a couple minutes and a few commands. It blows my mind that nobody else has gained any critical mass to compete here. Worse yet, though, is my disappointment that so many Operations teams (or DevOps teams, if you prefer) haven’t adopted a more services-oriented approach to serving their internal customers. Why do so many developers still have to request resources as if they’re physically allocated by a black-tee-shirt-wearing neckbeard sitting behind a CRT, rather than targeting internal provisioning APIs?
Q: Any tools/services that you think are totally awesome and feel more people need to know about?
I love seeing the explosion of tools within the monitoring ecosystem, both open source and commercial. Four years ago I wrote about this notion of “Voltron” where we had this composable system of small, sharp tools, and it’s really come to fruition over the last couple years. At the time, there were very few tools commercial tools that were examples of this philosophy, but now it seems that everyone has covered some portion of the monitoring landscape as a hosted service. Even better, I love that so many of them tie together well; integrations has become a huge focus in terms of competition and feature parity.
Q: What conferences or events are you looking forward to attending the remainder of this year?
This has been a very quiet year for me in terms of conferences. Outside of Monitorama in Portland, I’m heading to re:Invent in Las Vegas to help with our booth. I really would’ve liked to attend GopherCon just to see the energy and excitement in that community (although I generally dislike language or product-specific events), or perhaps MesosCon to see how companies like Twitter and Apple are using that technology.
I’m really excited for AutomaCon in Portland. I love the people at Heavy Water Ops that run the event, and I’m sure they’re going to put on a fantastic show. I’ve long since considered running my own systems/infra-as-code type conference, so their program really resonates with me.
Q: Is there anyone in the DevOps (or IT space) you would love to connect with, but haven’t so far?
Well, I’m in the fortunate position of running a fairly successful conference series, so I’ve gotten to meet or host some very amazing people. I think it would be great to spend more time with many of them, especially folks like John Allspaw, Bridget Kromhout, and James Turnbull.
I think, instead, that I wish everyone else would get to meet some of the fantastic people I’ve gotten to work with over the years but you might not know as well because they largely work behind the scenes, helping others be successful, while just generally being amazing human beings. Folks like Michael Gorsuch, Chrissie Brodigan, and Will Maier spring to mind immediately, but I’m sure I’m missing so many others and I apologize for that in advance. 🙂
Q: How or where do you consume the latest ideas and topics regarding DevOps and/or Agile?
I doubt there’s anyone left that isn’t subscribed to it yet, but if there is, I highly encourage you to join the Devops Weekly newsletter. Gareth Rushgrove puts together a very rich set of resources every week, without fail. It’s an impressive task to crank something like that out continuously.
Q: What’s the biggest pain point you plan to put off as long as possible?
Wow, maybe switching from Mutt to a graphical MUA? Every couple years I get fed up with having to read and write HTML emails so I consider moving back to Mail.app or Gmail’s UI, or whatever hot-email-startup happens to be headlining Hacker News that week. Sure enough, I keep going back to Mutt because everything else is “terribad”.
Q: From all of the main DevOps topics, which one(s) take up most of your cycles?
Monitoring, hands down. It’s the one thing I’ve always been passionate about, even before I realized it myself. I love building tools to help others be productive, and I really get a kick out of helping folks construct visualizations that transform raw data into something with actionable insights.
Q: Do you have any good books or resources that folks just starting to explore DevOps should check out?
Honestly, I would try and follow John Allspaw, Mark Imbriaco, and Lindsay Holmwood on Twitter and ask for their suggestions. These three are some big thinkers when it comes to systems and how we humans interact with them.
Aside from that, I’d love it for readers to check out my book on Graphite monitoring. Even if you’re using a different time-series stack, I think there’s a lot of really useful insights in there, especially as they pertain to composable monitoring architectures.
If that doesn’t sound like your cup of tea (or even if it does), you should check out James Turnbull’s upcoming book on monitoring. We’ve had a lot of really good conversations on these topics and I love his message on push-vs-pull systems.
If you currently leverage Librato for real-time cloud monitoring, I highly suggest you look at theLibrato/VictorOps integration to manage not only the alerting step of the incident lifecycle, but all phases and aspects of being on-call.