754 sats \ 4 replies \ @WeAreAllSatoshi 8 Sep 2023 \ on: First draft of my "Why you shouldn't use Stacker News" TABConf slide deck meta
From a tech stack perspective, I’d argue that no microservices is actually a good thing right now. It allows you to iterate faster. For every feature that requires DB, API, and UI changes, you’d have to coordinate multiple deployments to release. Maybe at some point it will be worth it to explore microservices, but I personally don’t think it’s a negative right now
Yep. Over engineering before you need it is a big mistake I've seen over the years. When starting a new project the most likely issue is no one will use it. Not that it won't scale.
reply
I agree. I read somewhere that microservices are not a form to organize code but teams.
That made a lot of sense to me.
I also got quite burned at my previous company with microservices.
Feels like the promise of "everyone can just work on their own projects" + "we can scale better" turned into "no one knows what everyone else is doing" + "if one service breaks, everything else may break in unknown ways because no service is really independent and the one guy who knows how to fix is on vacation"
Currently, only the image proxy is a microservice. This makes sense imo.
Basically everything that would be a container anyway can be turned into a microservice without drawbacks (?)
I know I am probably selling microservices short here
reply
I also agree. This may be a caricature of industry standards. Still, this is something I can imagine an "industry expert" saying so I'm not sure if I should axe it.
reply
I think it wasn't about axing it, just personal opinions here :)
same for "no typescript"
reply