Almost all organizations I see embarking on the platform engineering journey fall into the same traps. They find their way through them eventually, but I think we can save quite a bit of time by sharing these fallacies with everyone. So I started to jot them down in an article called “The top 10 fallacies of platform engineering”. I’ll expand on this in the future and maybe even write deep dives for each fallacy.
The first one is the most obvious and the first thing you’re likely getting wrong today: the prioritization fallacy. And you’ll understand instantly what I mean if you’re following this thought experiment: think of the entire lifetime of a developer, now your job is to optimize efficiency and experience. Where do you start?
90% of you will start with the onboarding of the developer. Why? Because that’s intuitively what you think of first. But from a strict ROI point of view, starting there is nonsense. If you look at the total lifetime of a developer, the onboarding process is below 1%. So even if you bring this down by 80%, the overall impact is minimal.
So where do you start? You list all the things your developers do that go beyond the simple update of an image. You calculate how often they do them out of a hundred deployments. Then you attach the times that devs and ops spend on them. And you strictly prioritize based on this.
Another example is the visualization fallacy. People think that by just visualizing stuff, it gets solved. Bullshit. My buddy Lee wrote a great piece called “Why putting a pane of glass on a pile of sh*t doesn’t solve your problem”. Worth reading.
Also, a couple of cool webinars coming up, check them out.
- The CEO of Logilica discusses what key DevOps metrics you should look at to improve your engineering setup
- The founder of Lightstep shows how to enforce observability for cloud-native setups
- The VP Product at LeanIX compares service catalogs vs IDPs
Stay relaxed and northern hemisphere people: spring is ney, maybe!
Kaspar
‍
