DevOpsDays Chicago had its first event in 2014, and 6 years later, this event is still going strong with about 500 people in attendance at the UIC Forum in downtown Chicago.
Matt Stratton and Sasha Rosenbaum started this year’s event by going over some of the history and evolution of the conference. Jeff Smith kicked off the day with the first talk: Ethics in Computing. Jeff proposed the idea that he’s not here to bring answers to ethical dilemmas in computing — but more to begin the discussion. Jeff presented many instances of these ethical issues in computing. He reminded us of the failures of companies like Amazon and Facebook to protect your data but also concerns about how they might be using it. In the end, it was the perfect talk to set up the event and really helped to guide discussions in some of the engaging open space sessions later.
Next up, Heidi Waterhouse moved into more of a technical discussion on software development and its evolution in the modern open-source world. The rise of microservices makes it even more challenging (if not impossible) to thoroughly test a complete system. Because of that, engineers need to focus on decoupling their testing to test endpoints instead. The main focus of the talk is stressing the point that folks need to kill their staging sites — and instead test in production. By leveraging techniques like feature flags for dark shipping services, you can move faster while still ensuring the ability to test features on the actual production systems.
After the break, Kevin Harriss talked about how to take an evolutionary approach to DevOps. He gave us great insights into how his company Enova took small steps to effect change in their enterprise. Kevin did a fantastic job detailing out how they were able to reduce risk and improve how work was able to get done. It was an engaging DevOps transformation story with some excellent real-world examples of steps they undertook. First, they followed a familiar pattern that many companies try, which is to create a DevOps team. As the scope of the team grew, they found themselves quickly overwhelmed by incoming work. Instead of scaling out the team, they moved to a “platform team” model, where dedicated groups of engineers would be tasked with specific platform deliverable projects. That way, each team would be following DevOps practices versus having one silo getting overloaded with work.
Lakshmi Baskaran closed out the first day of talks sharing the successes and failures of a DevOps transformation within an enterprise organization. She said that you can’t treat a transformation as a project with a clearly defined start and end date — it’s a continual journey for the business. I loved that to effect change in the organization, she found and promoted internal DevOps influencers who could implement and share successes.
The afternoon continued with some fantastic ignite talks (5 minutes long with slides that advance every 30 seconds), as well as many engaging open-space sessions. One particular open-space session that I love to help run at DevOpsDays is the #TalkPay session. J. Paul Reed started doing this at DevOpsDays Austin to help with pay equality. I’ve had the pleasure of running a #TalkPay with J. Paul Reed at a handful of other DevOpsDays events, and it was great to do it again in Chicago.
I was thrilled to lead off day 2 of the talks, where I discussed how to retain the hardest to hire engineers. Many companies spend countless time and money recruiting engineers, but then wholly fall down when it comes to keeping them. I went through some learnings over the past many years of building engineering teams and shared some successes and failures.
After my talk, Holly Allen gave a fantastic talk about service ownership at Slack. It was very cool to hear about her time working as a hardware designer and what testing looked like when you go from a hardware design spec to physical devices. She then continued on talking about how Slack moved from a traditional centralized operations team to having service owners manage their products and applications. It’s a model that many companies are moving to as they grow and scale. The main reason that Slack made the change was due to the number of products being built was far exceeding the ability for a single ops team to monitor and maintain. What I really loved about this talk was Holly taking us through the story, step by step, sharing all the good and bad that comes with significant organizational changes. The last two talks rounded out the morning presentations. Baron Schwartz talked about how to bring DevOps to the database, while Karissa Peth from Microsoft gave an engaging presentation about how to improve our documentation by learning from the scientific community.
The organizers did a fantastic job creating a wonderfully inclusive event with a diverse set of speakers. The open space topics in the afternoons were equally engaging with lots of different options for the attendees. It was great to see old friends and meet new folks and learn about some of the challenges they are facing in their roles.
See you all next year!