I Rewrote My Back End in Go in 72 Hours and Cut My AWS Bill by 94%

odilelof 55 points 24 comments September 08, 2026
www.bajura.online · View on Hacker News

Discussion Highlights (13 comments)

odilelof

Tuesday, October 3rd, 9:17 AM. I'm sipping my coffee, opening the AWS console like I do every morning. Except this time, I almost spat it all over my keyboard.

abdullah2993

You never mentioned what is this BE for?

pjmlp

Well, another one discovering strongly typed compiled languages. Apparently Tiger Beetle also started in node, before the founder reached out for Zig. Meanwhile all JavaScript tooling is being rewritten in Go and Rust. Maybe it is time to finally grasp it should be left to the browser?

mywittyname

Did you keep RDS Postgres?

exabrial

A rewrite is always better... I bet you could rewrite in bash with fastcgi and get an improvement

saejox

another good target for ai rewrite: getting out of serverless cost hell. it used to be a gigantic endeavor. now a few days work.

WesolyKubeczek

What stood out for me was that a complete rewrite of the backend application took 72 hours. Even pulling all-nighters. Especially pulling all-nighters. If I did it even 20 years ago, back when I was young, I'd expect the quality of my code to nosedive. If it indeed took this little time, I strongly suspect the backend wasn't doing too much in the first place. Usually you can rewrite those quickly and demonstrate speedups. Some real big software that had a misfortune to have been written in something interpreted and inferior, might be more problematic, especially if you can't afford downtime or are unsure about all important corner cases, and when the program itself is your only documentation. Another thing that was very jarring was that the article as a whole sounded very LLMese. Or PowerPointy? Take your pick. This little example screams that an LLM might likely have writen it: > I'm not going to give you a lecture. Here are the concrete numbers: > RAM — the real deal > > Node.js, on my latest version, was using 1.4 GB of RAM just to boot the app on one instance. > > Go? 48 MB. I'm not kidding. When talking to LLMs, I cannot now unsee how often they feel a need to assert something is real and they are not making things up. When I see this much LLM-like writing, I'm asking myself, maybe this all is made up? A plausible story, but maybe it never happened? Why would I spend my time reading it?

nhumrich

The real cost savings was from realizing the infrastructure needing nuking. Probably could have gotten almost the same savings keeping it in node and refactoring. Go doesn't seem to be the real hero of this story.

CodeWriter23

Looks like you hit paste 2x bro. : )

jezek2

Good, now you can cut your bill again by 94% by renting a normal VPS with unlimited traffic. The traffic is about 100x more expensive on "the cloud" and it killed all the ideas I had for cloud usage. Clouds are not worth it, even for scaling you can always rent more VPSes (or dedicated servers for heavy projects).

peter_d_sherman

>"Why Go saved my a* I'm not going to give you a lecture. Here are the concrete numbers: RAM — the real deal Node.js , on my latest version, was using 1.4 GB of RAM just to boot the app on one instance. Go? 48 MB . I'm not kidding. On the first test I ran, I launched a 15 MB Go binary on a t3.micro. The API responded in 20ms. I thought I'd measured wrong. I checked three times. It wasn't a mistake." There is definitely something to be said for a self-contained high-performance server in 48 MB...

mitxela

This is AI slop

chmod775

In a sane world this would still pass as satire. In our world I'm not sure whether a human even read the entire thing before publishing it. There's absolutely nothing of substance in here.

Semantic search powered by Rivestack pgvector
5,917 stories · 53,755 chunks indexed