The Velocity conference ends its run in Santa Clara, Calif., today amid ongoing conversations on everything from containers to microservices. We’ll be devoting entire blog posts to several key sessions at the event in the coming days, but I wanted to focus on some of Thursday’s highlights right away.
Engineering for the long game
The tech world is often dominated by short-term concerns, but in one of the morning keynotes, Google’s Astrid Atkinson explained how to take a longer perspective. “Imagine you will be very successful,” she said, and “act as if you will have to live with the consequences [of your actions] for a long time.”
She shared plenty of details from her years at Google, but I came away with these dozen takeaways that seem to have broad applications:
- Too much infrastructure is almost as bad as too little.
- You have to manage complexity, you can’t roll it back.
- “If you’re not careful, you can end up just moving complexity around.”
- The team is your most important asset, do everything you can to keep them with you. “It’s hard to manage this if you’re dealing with a lot of churn.”
- The team can’t grow as fast as the demands on it, or you end up with a really big team.
- You job is to not limit growth; you want to enable all the growth the product can drive.
- Strive for as much standardization as you can get. Anything you can standardize is going to pay off in spades.
- Ask what you can do to make things cheaper.
- “I strongly endorse boring infrastructure choices”—the most stable is probably the best choice.
- Ideally, a shared system should be self service. You don’t want to open a ticket for every change. “That’s a good place to take a hit on the dev side” for a long-term benefit.
- Don’t get wound up in the details. Pay attention to what outcome you’re trying to create.
- There’s no finish line to this. You can’t just declare victory and walk away.
Watch Astrid Atkinson’s full keynote speech here.
“Empathy” has become a hot, if sometimes fuzzy, buzzword, but author and speaker Indi Young took a practical approach, sharing tips on how to use empathy to help generate more and better ideas in the workplace. Young noted that “ideas pop up when you’re not actually trying to solve the problem,“ and that bringing in more people can “vastly increase the number of ideas in play.” In addition to bringing their own ideas, adding more people to the process also “vastly increases your own creativity.” In addition, it can help validate ideas you may not be sure about, she said.
And empathy—particularly what she called “cognitive empathy”—is the mechanism to do all this in a deeper way. Most people, she said, apply empathy based on guesses, often built on the demographics of the person involved. That doesn’t work, she said, merely reinforcing your own preconceptions. The key, not surprisingly, is to actually listen to what people say when you ask them.
Watch Indy Young’s full keynote speech here.
Patrick Lightbody interviewed on video
Patrick Lightbody, New Relic’s VP of product management, was interviewed about the New Relic Software Analytics Platform. He also discussed the challenges around the cloud, containers, and Platform-as-a-Service, and the emerging trends we can expect to see take center stage at Velocity 2016.
Docker book signing draws a crowd
New Relic’s Karl Matthias and Sean Kane drew a long line of readers hoping to score a signed early release copy of their upcoming book: Docker Up and Running. Sean and Karl graciously took the time to discuss containerization and other important topics with each person, making the wait more than worthwhile.
Data Nerd Sessions are standing room only
The informal Data Nerd Sessions at the New Relic booth featured cushy chairs for attendees, but pulled standing-room-only crowds all day long. From Stevan Arychuk’s “Taming Microservice Sprawl with Service Maps” to Nate Heinrich’s “Taking the Pain Out of Alerting,” they were all well received, but perhaps my favorite session was the last one: Bill Kayser’s “Three Performance Anti-Patterns: Mis-Indexes, N+1 Query Problems, and Cascading Failures.” As you can see, the crowd spilled out of the booth entirely.
The importance of design in ops
Finally, OmniTI’s Robert Treat closed the day with a session on What DevOps Can Learn From Design. His engaging talk covered the role design plays in many aspects of our lives, and concluded with a single, simple idea that’s worth repeating:
“We often claim incidents are caused by human error, but in many cases the human error is a result of poor design.” We would do well, Treat suggested, to make it a little bit harder to do destructive things.” Thinking about problems with a design-centered approach, he added, gives us the best chance for improving our world. After all, “There are enough things that will go wrong that are outside your control.”
Words to live by.
See tweets from other Velocity attendees here!