2012-06-29 16:23:23 by jdixon
This week I traveled out west for my first Velocity conference as an attendee. I went out two years ago but I was so busy juggling exhibitor duties that I didn't get to enjoy any hallway networking or formal session. This year I went in with plans to catch as many sessions as possible, particularly those skewed towards monitoring, trending and operations workflow. As expected, I skipped quite a few talks but made up for it with a lot of quality time catching up with peers and reviewing new technologies (and philosophies) in the DevOps space.
Early on I caught up with Theo Schlossnagle (my old boss from OmniTI and Circonus). He showed me Circonus' new histogram and heat-map features. Very cool stuff, nice to see this added to their already-exhaustive trending capabilities. In fact, the charts were being used on TVs around the Velocity hallways to demonstrate the conference wifi activity.
Finally met with Michael Leinartas in person and discussed roadmap possibilities for Graphite. A lot of our recent conversations have centered around a simplified web service option as well as decoupling carbon's dependency on graphite-web for reading metrics off disk.
Al Tobey reached out to me to demonstrate a new time-series database he's been working on. It's an impressive piece of work, written in Ruby and lots of good comments. Pros: it uses Cassandra for metric storage and has an extensive API. Cons: a lack of transformative functions and no built-in visualization tools. Al mentioned that he would like to port Tasseo to support it, which would be cool. The project has not been officially announced yet but I'll probably take a closer look at it when this happens.
Walked around the exhibition booth to check out the state of commercial monitoring software. There was a particularly nice monitoring dashboard from AppFirst. Unfortunately it's yet another push-monitoring service that hosts your data in their walled garden. I had a lengthy discussion with one of their engineers, making my case for decoupling monolithic monitoring apps into composable services.
Patrick Debois introduced me to his boss at Atlassian, Matt Moor. Matt happened to be talking shop with Adrian Cockcroft. The conversation diverged into a very interesting overview of Netflix's internal visualization tooling and some projects they have been meaning to open source. Of particular interest, Adrian described Atlas, their metrics storage engine, and some of their work on intuitive thresholds and alerting. Adrian was kind enough to Invite me over to Netflix to see some of their tools in action.
In my spare time I continued work on a new dashboard project (Descartes). Where Tasseo is designed for heartbeat monitoring on NOC-style large displays, Descartes is intended for deep-diving and collaboration. I hope to have something ready for public review by the end of next week.
All in all it was a fantastic week of information absorption and bouncing ideas off my contemporaries. I'm pleased to see the #monitoringsucks movement progressing and that so many people are receptive to the concepts I'm "dialing in". Even the flight home was pleasant and surprisingly engrossing. On the other hand, perhaps I need to take a break when everything I see looks like a chart.