Monitorama 2017: The Monitoring Revolution Continues

Once again, Monitorama 2017 — the sixth official Monitorama — was held in Portland, OR. The event began at 10 a.m. on Monday, May 22 with a talk by John Rauser of Snapchat and ran for three days packed with great presentations, demos, and conversations.

Day-by-Day

John kicked things off by speaking about the future of the monitoring toolchain. One of his lines — “You’re all data scientists; you just don’t know it yet” — hit home with the concept that many of us are at the forefront of the monitoring revolution. He continued by taking traditionally difficult, high-level monitoring ideas and distilling them down in a way that can be understood by the masses. In particular, he did a great job showing how the R statistical language, specifically the Tidyverse suite of packages, is much more flexible than what some of our modern data visualization tools provide.

Alice Goldfuss, an engineer at Github, followed up by speaking about the necessity of being on-call, and the beginnings of the #oncallselfie while she was working at New Relic. She shared her on-call experiences, which definitely struck a chord with the large group of operations professionals. Her main point was that oftentimes you can’t tell whether you’re getting “too many pages” or “being paged too often.” And the best way to improve your situation is to raise your hand to let others know.

The rest of Day 1 was filled with great examples of how companies actually “do” monitoring. Netflix, Slack, Capital One, and many others showed off their tools and tips of the trade.

The first day also featured “The Vasa: Redux”, my presentation that explores the parallels between the failure of the Vasa and how software engineering projects fail. If you’re interested, you can view it at: https://www.youtube.com/watch?v=P8dc-rLnLr0&t=8h6m00s

PeteVasaMonitorama.png

Day 2 started with a total venue change due to an underground fire that knocked out power to much of the western part of downtown Portland. If there is one thing Operations people thrive on, it’s dealing with failure. The organizers did an incredible job of relocating all 650 people to a new venue, and the show went on without missing a beat!

Day 3 began with a fantastic talk from Travis CI engineer (and Threat Stack customer), Brandon Burton. Brandon talked about the challenge of monitoring systems where you can’t fix the problems. Travis CI manages the continuous integration and build systems for many open source projects on donated hardware, and Brandon discussed the challenges of getting useful data from the build logs to provide actionable info to their users.

Amy Nguyen, software engineer on the Visibility Team at Pinterest, also gave a talk that resonated with the audience. She discussed how important UX is when building monitoring tools for your users, sharing her experiences at Pinterest and showing off some of the great internal tools Pinterest has built over the years.

For video highlights of all three days, check out:

Reflections & Observations

It’s been great to watch the conference evolve over the years. New topics are constantly being introduced to keep pace with the changes and challenges in our industry. (Distributed tracing, for example, was a popular topic for speakers this year, as more and more companies embrace microservices in their organizations and need a reliable way to understand latency across many distributed applications.)

But one of the best reasons for attending Monitorama continues to be its unique Single Track conference style (which means there’s no fear of missing out because of scheduling conflicts) and the extremely dynamic hallway track. Sharing experiences and ideas with your peers during breaks is one of the main reasons I continue to attend Monitorama and why Threat Stack continues to sponsor.

Key Takeaways . . .

All in all, it’s great to see such a passionate group of engineers getting together to further level up their trade and share their experiences with a growing group of people who are interested in making monitoring better for all.

I don’t consider Monitorama to be part of the DevOps space (mostly because I hate the term DevOps in general). HOWEVER, since successful DevOps means understanding monitoring, then yeah, Monitorama is a very unique event in the current space of technical conferences.

It stands alone with its focus on talks from engineers and operators who want to improve our craft of monitoring. Before Monitorama started, the common line was “monitoring sucks” — but as the use of newer technologies like AWS and other SaaS services came about, the need to improve the tooling to handle this new world became a critical issue.

And that’s what Monitorama does — it provides a venue for people to show and talk about how they solve problems using modern monitoring tools. The conference talks and workshops drive attendees to further improve the tools they are using and contribute back to community at large, continually pushing the boundaries of monitoring and trending software.

See you next year at Monitorama 2018.