Many of the videos from KubeCon EU are on YouTube now, just days after the the talks were given. That’s pretty great! Here’s three talks from the Platform Engineering Day that I really liked. They’re primarily focused on what I’d call designing and product managing a platform. I’m a fan of the “platform as a product” notion - I came across it around 2016 or so at Pivotal and saw many organizations apply it. These three videos are in that general area: how do you build a good product that application developers will want to use and “get value” out of, not just deliver IT capabilities and services.
They’re worth watching all the way through:
“Sometimes, Lipstick Is Exactly What a Pig Needs!”Abby Bangser, Syntasso & Whitney Lee, VMware - How can you start to take a design/UX-driven approach to building your platform (you know, your pile of Kubernetes stuff to pull it all together for app developers)? Whitney Lee & Abby Bangser have a good mind-model to think about designing the interfaces (how people use parts of that stack). Thinking through how you design your platform “interfaces” for app developers is especially important if you’re building your own stack instead of buying one of the already integrated ones.
“Designing for Success: UX Principles for Internal Developer Platforms,” Kirsten Schwarzer, KubeCon EU, March, 2024. - Good tour of applying design-think to platforms. That is, what do designers do on platform engineering teams. Some items: (1) Do at least one hour long user interview a week. This gives you an idea of what your users (developers) are doing, struggle with, how your improvements are helping or not, and give you data to decide what to do next. The last is especially useful for fending off The Boss and others who have suggestions that are not helpful: show them the data about what is, at least that you do things driven by analysis, not hunch. (2) Progressive disclosure - only show what people need for a given talk, not everything the tool can do. Can help with cognitive (over)load. (3) Errors are a good source for finding what’s wrong.
“Boosting Developer Platform Teams with Product Thinking,” Samantha Coffman, KubeCon EU, March 2024. - A lot of good commentary and advice on product managing the platform, some techniques, mindsets, etc. Chief among them: focus on the value/problem you’re solving first, then how you solve it second…also trying to quantify what the problem is so you can measure if it’s fixed, and prioritize it.
With platform engineering, we’re asking operations and infrastructure people to think differently about what they do and how they do it. We’re asking them to think in terms of products, design, customers. This is in contrast to the traditional/current way of thinking about infrastructure which is more like: you tell me what capablites, capacity, and uptime/reliability you need, and I will deliver it to you. What you do with all that is not my concern. This is fine!
When it comes to a platform, though, the platform engineering people are want the platform team to solve problems for developers, not just give them the foundation needed to run their apps. These are different concepts, skills, tasks, and ways of thinking than existing infrastructure think and many won’t know exactly what to do. The talks above, then, are good at walking through new ways to think about “delivering” infrastructure, new ways to evaluate and plan for how you stack things together and gather up requirements.
There’s more videos to check out from the little platform conference, and a ton from all KubeCon, of course.
Oh, and also: If you’re thinking about building and running platforms. You’re probably thinking of just piecing together your own out of free parts and little doo-dads here and there. That usually doesn’t work well. We have a fine selection of pre-shaved yaks you should try instead.
Check out the video series here.
“The Old Town is Old and the New Town is … well, less Old.” A correspondent on Edinburgh.
“Grandboss” and “great-great-grandboss.” Here.
“Something I’ve been thinking about lately.” Here.
“Build something cool that also meets my company’s requirements for security.” Here.
There’s not always an “always,” even if there’s often a “usually.” Ibid..
Talks I’m giving, places I’ll be, and other plans.
Tanzu (Re)defined online, April 3rd, The Internet. Tanzu (Re)defined, April 11th, Palo Alto. TEQNation, May 22nd, Utrecht, speaking. DevOpsDays Amsterdam, June 20th, speaking. NDC Oslo, speaking, June 12th.
In case you didn’t notice it, if you’re interested in what Tanzu (where I work) is all about, we have an online event now, on April 3rd. It’d be cooler to go to the in-person on on April 11th, but it’d be coolest if you make it to at least one of them. The online one will be broadcast in YouTube and LinkedIn. Sign yourself up for it, why not?