Earlier this week I had the pleasure of attending DevOpsDays Rockies at the Mile High Station and Iron Works in Denver, Colorado. This was a two day conference focused on the intersection between software development and IT operations. The conference started off on a fun foot when I found myself along with another person stuck on a freeway overpass overlooking the event venue thanks to some poor Google Maps walking directions. As it so happened, the person I was stuck with was J. Paul Reed, the keynote speaker for the event. After some additional walking and a ride from one of the event organizers, we successfully made our way to the conference.
J. Paul Reed's keynote focused on the human side of site reliability engineering (SRE): how incident response could be improved by factoring in behavioral heuristics. The talk covered fundamental SRE questions, such as How do you know when an incident is occurring? and How do you know how to respond to an incident? But did so with a different approach than I have seen in past talks. He dove into concepts like system 1 vs. system 2 decision making, the Rasmussen Triangle, and how directed, manufactured, and vicarious experience could be used to develop expertise in order to resolve incidents. The content of the talk was both useful and not something I would have expected from an IT conference.
One of my favorite parts of the event was Open Spaces. If you are unfamiliar with the concept, Open Spaces provide a forum for conference attendees to gather and discuss topics outside of the main tracks. Attendees propose topics via post-it notes or another medium and then vote on the topics they would most like to discuss. These topics can range from specific technical concepts, such as the best approach to service discovery in a container native environment, to compensation and how to convince upper management to adopt a DevOps methodology. Because of the diverse experiences brought by attendees, these forums are a great way to learn more about what other people in the industry are doing to solve particular problems. If nothing else, Open Spaces are a great excuse to start conversations with other attendees.
On the first day I attended an Open Space simply called "Jenkins X" and another one comparing service meshes. I had some familiarity with Jenkins X, but I was hoping that by attending the Open Space I could learn even more. While most of the crowd was aware of Jenkins, very few had used the newer version of it designed specifically for Kubernetes. As it so happened, I was one of only two people in the group of 20 who had any amount of hands on experience with Jenkins X, so rather than sitting back and absorbing as much material as possible, I ended up leading the session and sharing my experiences with the tool. The service mesh discussion was much more along the lines of what I expected; I was able to listen the the experiences of companies using service meshes, including Linkerd, Istio, and to a lesser degree Consul. One of the companies had been using service meshes in production for years. One of the most interesting takeaways from that talk was that most of the people in the group used service meshes for traffic management and network routing rather than for observability.
On the second day I proposed a couple of talks myself, one of which, a comparison of continuous integration and continuous delivery (CI/CD) tools for containerized environments, ended up being the top vote-getter *pats self on back*. My goal for the topic was to learn more about the pros and cons of CI/CD tools used with containers. Similar to the previous day, the experience of the group was limited in this area. No one mentioned Spinnaker and no one had hands-on experience with Jenkins X. Travis CI and Circle CI + Argo were also brought up as options, but people had merely experimented with them and no one was using them in production. I was surprised to learn that a number of people were using GitLab in production for this specific use case.
One aspect of the event I appreciated was an attention to sustainability. Rather than distribute the perfunctory conference t-shirt, DevOpsDays handed out reusable drink containers and bags. They also made a point of mentioning that while single-use cups were available for coffee, the containers we received were suitable receptacles for warm drinks. This was a great method of reducing waste produced by the 500+ conference attendees. Another aspect of the event that I appreciated was its inclusivity. The last two DevOpsDays I have attended have provided gender neutral or non-binary bathroom options and made an effort to use inclusive language throughout the event. DevOpsDays Rockies also had a line up of speakers that featured more than 60% women.
Oracle Cloud Native Labs was very pleased to be the coffee sponsor for this event, and we look forward to participating in more DevOpsDays in the near future.