Post

I genuinely enjoyed reading this. I have additional thoughts on the loop. 1. Rapid feedback and iteration necessitates not just observability but fundamentally different test tooling that treats test as simulation, not box checking.

Charity Majors@charity.wtf

I've come to view the entire DevOps movement as a mighty, twenty-year battle to achieve one thing: a single feedback loop connecting devs with prod. It failed, but not for lack of trying. It failed because the technology wasn't good enough. And now it is. www.honeycomb.io/blog/you-had...

honeycomb.io“You Had One Job”: Why Twenty Years of DevOps Has Failed to Do itLet’s start with a question. What is DevOps all about?

30/1/26, 5:39 pm0 reposts9 likes

Ada Lündhé

@adalundhe.dev

30 Jan

2. Such tooling needs to be “write once, scale anywhere”. The simulation needs to scale and work whether running on an engineer’s laptop or a sophisticated multi-region deployment.

Ada Lündhé

@adalundhe.dev

30 Jan

3. The results and insight need to be dimension rich but carefully selected, customizable by engineers to gather the information *they* need to act, and work with/enhance standards like OTel.

Ada Lündhé

@adalundhe.dev

30 Jan

4. The tooling needs to facilitate more than rapid scaling of linear happy paths - it needs to be able to ingest and utilize past metrics around usage patterns, traffic levels, etc.

Ada Lündhé

@adalundhe.dev

30 Jan

More than anything, it needs to feel actually *nice to use*. So much of observability tooling and test tooling *still* has gnarly, very sharp edges.