Very relatable. I end up stressing myself out over my side projects until I remember to be kind to myself and that I’m only doing it for the joy of coding!

I’m at the ~3mo mark of learning and using Neovim and I love it, but I have had a few cases where it hasn’t worked well and I’ve fallen back to VSCode (mainly for C#). So I’ve been debating whether I should just stick to the tool I already know.

Great to hear your perseverance has paid off!

“The cup’s getting full, can we extend by a TB?”

Save us please

Just enjoyed reading www.principalengineer.com/p/old-school....

I've been thinking a similar thing lately, and have liked using Astro's Islands concept to write a 'good old server-rendered' app with bits of React for pockets of interactivity.

Not thrilled about JS on the server but I'm having fun.

PS @lethain.com I really enjoy your writing - keep it coming please 😀

Good reminder that technology decisions aren’t pure functions - there are human side effects!

No I agree. Could we have used FaaS? Probably, and there probably would have been benefits. One of the goals was to increase supportability. The team supporting the system day to day were generally not developers, so the abstraction/misdirection with FaaS was unhelpful.

I really wish Dapr’s workflow engine had existed then. It would have saved us from loads of the workflow machinery.

docs.dapr.io/developing-a...

We did exactly that for an enterprise customer’s integration services last year - stateless containers in AKS (Azure EKS equiv.) with KEDA’s queue length autoscaling.

It helped that the customer had an ops team ready to learn and take on management, but they’ve not had any major headaches.