So I just came back from the sunny Barcelona yesterday, and I have to admit, Barcelona is az amazingly beautyful city. But more importantly Velocity is an amazing conference. This post will be a short recap from my point of view grouping my key takeaways from that conference.

Open source tools:

So my favourite topic, I’ve heard great talks about tools we already know and love, like Etsy’s very own nagios-herald, and some new ones i didn’t really hear about yet, but was totally amazed by, let this be just a short list, and check them out yourself:

Commercial tools:

I’ve seen some pretty awesome commercial tools, which were new for me and blew my mind as well. Let’s progress with a list which one is not odered, because i can’t decide order based on awesomeness.

  • Ruxit What Ruxit does is the following: It has an agent on your box, and monitors your system, this is really common for a monitoring system, right ? :) But the mindblowing little added value comes here: They call it contextual alerting. So, it creates correlations between incidents, so when you get hit by an alert you can see what was the root cause, what was the problem that originated from that roout cause, and the other little failing checks that originate from the same root as well. Saves lots of time for on call guys. Awesomesauce, right?
  • Cedexis Cedexis provides free R.U.M solution, and a paid Load balancing service. What Cedexis does, is they collect the network data in they free RUM tier, so they have a global map of the network traffic. And based on that, believe it or not, they provide the best routing between your service and the enduser. DNS load balancing based on global metrics, well this is my reaction:



We’ve heard lots of great talks around various technological topics, covering containers, anomaly detection, an particularly good talk by Google’s very own Ilya Grigorik about optimizing TLS performance. Just WOW.

Theoretical stuff:

As I mentioned before, we had great talks in the topic of anomaly detection, and oh boy, fiddling with time series comes with glorious amount of deep deep math, the talk by Arun Kejariwal was heavy on math, but I enjoyed every moment of it. The other anomaly detection talk was by the guy, who is highly recognizable by his ever changing facial hair, Theo Schlossnagle, the summary of his talk can be found here. We also heard pretty good amount of talks around microservices, deployment, and uptime monitoring.

Linux Containers

Linux containers are a really hot topic nowadays, and it’s totally understandable why. It’s the ops equivalent of the microservice oriented code. And as so, it’s meant to serve as a backbone of a microservice oriented infrastructure. There was a presentation, which one ripped apart the real meaning of linux containers, how they really work, and what docker, or LXC actually abstract away from the enduser. Slides The other great tutorial in this topic was about Core OS and Kubernetes.Material at Github


To summarize, I learned a lot, networked a lot, met lots of nice people, and enjoyed the show maximally. I’m really thinking of going back next year as well.


Long time, no see. Don't worry, I'm not dead, not even remotely dead. I'm just pretty busy. I'm actually working on great stuff at work and …… Continue reading

Openduty @ Cloud Budapest

Published on April 12, 2015

Leaving Ustream

Published on February 07, 2015