The tem “Velocity” has multiple, morphing meanings in the world of software development. For one thing, it’s an O’Reilly conference running this week in Silicon Valley—New Relic is there!—but it’s also a concept close to the hearts of just about everyone in the application development and management space: what we can do to make more stuff happen faster.
Both uses of the word come together in the tenth edition of The New Stack @ Scale podcast (player below), where I join host The New Stack founder and editor in chief Alex Williams in welcoming Tori Wieldt, New Relic developer advocate (and co-host of the New Relic Modern Software Podcast) and Charity Majors, co-founder of Honeycomb, a San Francisco startup working to “empower engineering teams to make decisions faster using real data.” Before helping to start Honeycomb, Charity formerly worked at Parse, which was acquired by Facebook and will be “retired” early next year.
Our freewheeling conversation touches on many topics, from the evolution of New Relic, to Honeycomb’s mission, to dev and ops learning to get along, to the rise and fall of Parse, to abstracting complexity. But I was particularly interested in a couple of areas:
Don’t just write code, watch code
First, I loved Charity’s answer to my question about how do you impart velocity to engineering teams [time code 10:43]. At Facebook, she said, the best engineers “spend less time writing code and more time watching the effects of it.… Part of this is because when you’re dealing with such a large codebase, the human mind cannot comprehend it. It cannot comprehend what ripple effect you’re going to have.” The key, she said, is that engineers have to get used to monitoring and instrumentation as not just “something you go to in time of emergency, but a thing that you use to augment your development process.… Writing code without watching your instrumentation and your visibility stuff is like a pilot who literally doesn’t have the tools to see where they’re going to land.… You have hundreds of thousands if not millions of metrics in a decently sized system. If you don’t know what things look like when it’s going well, then you don’t actually have the ability to spot outliers” when they occur.
Essential, but not taught well
That naturally flows to the second thing that stayed with me from our conversation: the importance of developers accepting responsibility for how their software works in the real world. Call it DevOps or whatever you like, it’s both essential and not properly taught and communicated.
When Charity talks to software engineers about the importance of operations, she tells them: “You shouldn’t feel bad nobody taught you this in college. The culture has not been teaching you this, your entire career, but let me tell you. You should want these skills. You should want this visibility into your code. You will blow people’s minds with how much more powerful you are when you know how to look at these metrics and understand what’s going on and you can predict problems before they happen.” [21:21]
Go fast, but know where you’re going
Finally, I thought Charity made a great point about the role of monitoring in modern software companies. “It is both absolutely critical to every single company and it is not the core mission of almost every single company.” [32:01]
Sometimes that dichotomy can tempt companies in search of velocity to skimp on monitoring. But going fast doesn’t always work out well when you can’t see where you’re going.
Check out the entire podcast here:
New Relic is a sponsor of the New Stack @ Scale Podcast. However, the content and views expressed are those of the participants of the New Stack @ Scale Podcast, which is the property of The New Stack. Any views expressed on the New Stack @ Scale Podcast do not necessarily reflect the views of New Relic. By embedding the audio for the New Stack @ Scale Podcast or linking to The New Stack, New Relic does not adopt, guarantee, approve or endorse the information, views or products available on The New Stack site.
Finally, for more insights into the podcast, read The New Stack @ Scale Podcast Episode 10: Platform Versus the Chaos Monkeys. Joab Jackson’s post on The New Stack focuses on Charity Major’s time as infrastructure tech lead at Parse, and how “developers can sometimes seem like living, breathing chaos monkeys, forever poking new holes into the infrastructure in enthusiastic, and sometimes utterly novel, ways” and how that can cause big issues at scale, such as when Parse became home to some 100,000 applications.