Monitorama 2018

After hearing how unique this conference is, I decided I needed to experience Monitorama for myself. I booked tickets, hopped on the world’s shortest flight (train didn’t leave early enough) and headed to Portland, Oregon.

Monitorama is a conference about … monitoring! I know, confusing, right? Well, my goal in attending the conference was to learn from industry experts about their paths and innovative approaches to effective monitoring.

Conference Overview

Monitorama took place from June 4th to 6th in Portland’s Pearl District. Surrounded by centers of collaboration (bars and coffeeshops), this conference provided plenty of opportunities to meet new people and discuss different ways to solve monitoring problems. The single-track sessions meant everyone hears every speaker. The normal flow was a two back-to-back speakers with a couple of sponsor talks in the middle followed by a break. Day 1 focused on softer/more abstract topics such as mentorship, alert fatigue and storytelling. Day 2 was a little more technical, diving into slack bots. Day 3 went in super deep, focusing on monitoring serverless applications as well as complex math for scaling up infrastructure

Conference Highlights

For starters, this was the first conference I’ve attended where Slack was used. A dedicated organization was opened up a couple of weeks before, and people immediately began creating useful channels such as “coffee-n-donuts.” All kidding aside, the Slack channel allowed everyone at the conference to collaborate, make plans and ask questions even before we all arrived. (At Liatrio, we generally don’t recommend going to conferences solo, but Slack would definitely make that easier.)

One question that always gets asked at conferences is what time will lunch will be served (and what’s on the menu)? Surprisingly, Monitorama had no catered lunch, which encouraged another unexpected development — the creation of “lunch parties” (coordinated in the #lunch-parties Slack channel).

On the second day, my coworker Shane and I coordinated a lunch party to discuss monitoring CI/CD pipelines. Seven of us attended, and it ended up being a great experience. Using the same Slack channel, Shane and I joined a great breakfast on day 3 right before the conference focused on tracing technologies. Just great.

The venue had fast internet which made keeping in contact with work, notes on google drive, and keeping up with the monitorama slack so much easier than it could have been. Did I mention how ideal the location was?

World class coffee is required when hosting a conference in Portland


On-boarding –

Three talks focused on onboarding for operations roles. It’s clearly a hot button topic. The most common rotation involves a buddy or assigned mentor who works with a new hire directly for a 6-week period with 3 distinct sections. Sections 1 and 3 had one person (mentee and then mentor) observing, while section 2 had them pairing on tasks.

Slack –

Slack clearly had exceptional adoption at the conference. It was used throughout the day by both speakers and attendees, and many of the speakers discussed slackbots they had written. Is Slack usage a sign of a high-performing engineering organization? Perhaps.

Serverless –

AWS Lambda was introduced in late 2014. At that time, it only supported Node.js and was understood by very few. Things have changed in 2018, where speakers at Monitorama seemed to see both Lambda and serverless platforms as the silver bullet our industry has been waiting for. No need to patch servers. Infinite scalability at the call of an API. Presenters called it a great tool for the backend of slackbots and other monitoring processes. And it’s not limited to internal tools. For example, one technical speaker described how to monitor serverless applications.

Observability over Monitoring –

Context was heavily emphasized in a number of talks. The story, or health of an action, is much more important than the health of a server. Hitting 100% CPU utilization can be fine depending on the circumstance, but it’s unacceptable for a user request to take too long or to fail.

@ajdomie  Every alert should come with context. Tools today can deliver a list of possible reasons for why this alert has shown up in the past.

Prometheus –

Prometheus is a very popular monitoring tool. When it was mentioned, people literally cheered. (One person gushed that it “can check anything that’s measurable”.)

@enbnt Monitoring tool progression at Twitter

For next time…

One thing that I felt was missing from this conference was a discussion of how developers can make their applications more monitorable. While a couple of presenters talked about the importance of writing better documentation for developers to read, I felt that a collaborative discussion was missing. In my experience, a majority of software problems are solved when operations and development teams work closely together.


Monitorama 2018 was a great conference. The topic of monitoring is an important one, and it’s an area where any type of engineering organization could improve. I will happily attend again in the future.

Leave a Reply

Your email address will not be published. Required fields are marked *