Thread by @QuinnyPig

There are a few things I take issue with in this report—like @monkchips’s use of the word “almost” in this section!

(Seriously, when someone like James says this about me it’s a tremendous honor.)

https://twitter.com/monkchips/status/1387419402673573892

If we go back in time three years, @SpotifyEng committed to spending $150 million a year on @GoogleCloud (assuming flatline / no growth. HAH!).

This report doesn't really highlight the stupendous scale of Spotify's cloud environment, so without further context I worry that it's going to be something that Twitter for Pets tries to adopt in its three person engineering team.

Don't do that!

I admit to being bearish on "in-house developer platforms." It makes sense at stupendous scale, but they're invariably one-company tools. While 40% of PRs to Backstage now come from non-Spotify people, I haven't yet seen it in the wild.

Backstage has been adopted by the @CloudNativeFdn, which is... well, it's not named "Kubernetes" so you can form your own opinions on that score.

Note that @monkchips mentions @SpotifyEng has over 500 teams. That's a lot! At this stage, all problems are inherently political / cultural. Trying to adopt what Spotify has done to your shop is almost certainly doomed to fail without a cultural transformation too.

I want to be very clear--Backstage is a marvel. So is Cost Insights. @monkchips has represented them fairly and well!

They just haven't seen breakout success yet; there are some adoption challenges they've gotta overcome before that's likely to happen.

One last point--@NetflixEng open sourced ICE years ago, and it languished. Then @teevity took over the ICE project. It said / did much of what Cost Insights says on the label.

The last commit was as of this moment 4 years ago.

github.com/Teevity/ice