Platform Engineering - Passing the Frustration Token Back to the Ops.

I'm on my way back from the unbelievably beautiful city of Prague where I had the honour of delivering a talk at KCD Czech & Slovak (together with Ivan Yurochko )

Besides the talk where we discussed testing #Kubernetes workloads for reliability and performance - conference organizers also put together a panel about Platform Engineering - led by Miro Adamy and featuring Jakub Stehlík , Zdenko Vrabel , Dotan Horovits ?????? and myself.

And was it an important conversation! I feel like we definitely touched a nerve there!

Trouble is - often when we talk about internal platforms it sounds like we're trying to convince ourselves they are actually needed. Especially so when it's @ doing the talking.

There are multiple blog posts and conference talks - all recycling the same notions. Platform Engineering is about developer experience, economies of scale, cognitive load, APIs, bla-bla-bla... It all makes sense, and yet - doesn't give us a good sense of direction. Is this actually a new discipline? How much should organizations invest in Platform Engineering? How do we measure the platform ROI? Are we gonna teach this in college? Should I recommend young people to become platform engineers? Or is this just DevOps engineers rebranding themselves to get a raise? (Something I jokingly said at the panel, but many people sincerely believe is true)

And here we were - five smart experienced IT professionals, trying yet again to make some sense in this buzzword soup.

I believe somewhere in this conversation there was a glimpse of truth. After all - Platform Engineering, just as DevOps before it, deals with humans trying to improve the way we collaborate in IT.

- "DevOps grew out of the Ops' frustration..." - said Zdenko Vrabel at some point.

- "DevOps practices, such as IaC and the whole shift-left movement has put too much cognitive load on developers." - continued Dotan Horovits ?????? .

And that's when it shone on me: the Ops were frustrated by Devs throwing changes over the wall, so they invented DevOps to slow them down. The intention was good, but now the Devs found themselves frustrated by all the "cognitive load". So they are pushing back, demanding better DevEx. Basically platform engineering is developers pushing their frustration back in the direction of the Ops! Or, as Miro put it - "It's all about passing the frustration token back and forth".

So yes, PE is about DevEx, APIs, delivery optimization and what not. But beyond all these there are frustrated humans trying to improve the way we work. And we can go a long way if we admit this frustration and realize that platform ROI isn't always measured in money, be it savings or profit. Sometimes it's measured in engineers' happiness. Life is too short to work with crappy tools, life is too short to suffer from inflexible processes, life is too short to be bored at work. Life is about creativity, open communication and delivering value. If you are doing PE - let it be first and foremost a frustration relief. And the money will follow.

I know this probably doesn't explain much either... But for me - it gives me the motivation to go on. I've always believed that happy engineers deliver great software. And in that respect - DevOps and Platform Engineering are the same. They are about making engineers happier. Hopefully not at the expense of other engineers. Or else - the frustration token will get pushed back yet again.