Last Friday I attended IncontroDevOps, one of the most prominent DevOps events in Italy.
It was a great opportunity to vocalise our thoughts about the human side of operations (checklists, incidents, postmortems, et cetera) and get some good discussions going.
Ready for war games #DevOps #IDI2016 pic.twitter.com/5lH3UzfSnR
— Marco Trincardi (@Trink0) April 1, 2016
IncontroDevOps took place in Bologna, home of the oldest university in the world and some pretty fancy supercars (Lamborghini and Maserati). It was a sold out event, with more than 200 attendees, sponsors and speakers in attendance.
That said, there was plenty to keep me busy. Here were some of my favourite tracks from the day:
#IDI2016 let’s start incontro DevOps Italia :-) pic.twitter.com/OTNAYy3h2R
— Roberto Bettazzoni (@bettazzoni) April 1, 2016
DevOps @ Ineca
The Ineca team (a non-profit consortium of 70+ education institutions) talked about their journey to DevOps. As soon as they started using Git, they realised how pull requests can bring together developers and sysadmins.
But they didn’t stop there. They also told us how they:
- Automated tasks with Jenkins
- Built docker images with Puppet, and
- Managed their container environment with DockerUI, Docker Swarm and OpenShift.
Check out their slides: DevOps @ Ineca.
Infrastructure as code
Yevgeniy Brikman demonstrated how to deploy a Sinatra / Rails stack on top of ECS (ECS is an AWS Docker service) using Terraform. It did look like the perfect development environment, but where it’s lacking is in its production capabilities:
- Updates: the Docker ecosystem still needs to provide a viable solution for software updates.
- Dynamic environments: how to deal with autoscaling systems, i.e. adding nodes as demand goes up, and deleting them when they are no longer necessary.
- Databases and big data: how to deal with persistent data on containers.
Check out the slides: Infrastructure as code: running microservices on AWS using Docker, Terraform, and ECS.
Puppet Evolutions
Alessandro Franceschi presented what’s next for Puppet4. Configuration automation tools are expanding into infrastructure automation. They are gaining the ability to configure network and storage devices, cloud infrastructure and containers alike (Docker, Kubernetes, and Mesos).
Here are the slides: Puppet evolutions.
Monitoring for Performance-Critical Applications
As a monitoring enthusiast I couldn’t help but enjoy Manan Bharara’s session on how Otto (one of the biggest German ecommerce websites) handles monitoring. According to Bharara, Otto run their own self hosted monitoring with Graphite. But instead of using the standard stack with Grafana, they use:
- Oscillator, for dashboards
- Xray for alerting
I was surprised they went the self-hosted route, but there maybe operational aspects I’m not aware of.
Slides: Necessary tooling and monitoring for performance critical applications.
Very cool work about continuous monitoring using open source tools by @mananbharara #idi2016 #devopsita pic.twitter.com/X24CatjNby
— Lucio Riccardi (@cantorjf) April 1, 2016
Web Performance in the World of HTTP 2.0
What I took from Dave Methvin’s session was that SPYDY will soon be deprecated. Most cloud providers and CDNs are ready to offer HTTP 2.0 and the latest versions of most client software already supports it. This might be the next big thing on Internet protocols (after IPv6).
Here are the slides: Web Performance in the World of HTTP 2.0.
War Games, Flight training for DevOps
In my session, I shared how we run War Games at Server Density. I demonstrated how the human factor has an impact on the incident time-to-resolution metrics. Our HumanOps strategy consists of the following practices:
- Using checklists to avoid human error
- Design on-call rotations to avoid fatigue
- Alert design, and response procedures
- Communication processes during outages
- Ops training, war games, and simulations
- Managing remote teams and work/life balance
Summary
IncontroDevOps was a well organised event with some very insightful talks. We look forward to next year’s event. Until then, we hope to see you in one of the upcoming DevOps conferences here in Europe.
Did you assist in IncontroDevOps? What did you learn? Let us know in the comments or message us @serverdensity.
The post What we learned at IncontroDevOps appeared first on Server Density Blog.