Fart Stack | Hacker News Reader

Fart Stack

fart-stack.io

107 pointstodsacerdoti36 comments

This made me laugh.

Just want to bitch: Anyone can get a kubernetes cluster up in an hour.

Very few people can get one back up that has blown up in an hour or can upgrade a cluster without downtime or can persuade terraform to do something without destroying the cluster due to drift.

They’re the sort of people that we need in the market.

There are dozens of us! Dozens!!!

But seriously, if anyone needs training on this (or emergency help!), shoot me a message.

I think the key here is drift. What is drift? Do you mean that the environment has shifted beyond what is defined by the terraform infra-definitions files?

How do you actually do something about this?

What causes it?

Humans. Infernal humans.

I can get Consul, Vault and Nomad clusters up in an hour though.

Can you fix raft when you get an election storm, unseal vault and have time left to rescue the princess in an hour though?

I get that this was made as a joke in addition to maybe a bit of self learning therapy; however, I would suggest that the author should add some prose to "How Does it Work" that explains how Kubernetes is involved in the stack otherwise as an employer wanting to hire someone I'm not convinced this person actually learned anything about Kubernetes with this project.

> otherwise as an employer wanting to hire someone I'm not convinced this person actually learned anything about Kubernetes with this project.

Not every piece of writing on the internet needs to be written to sell yourself to a potential employer.

> Right after I was laid off, I purchased a course on Udemy to teach myself Kubernetes.

Isn't that the point of the project though?

I tried to make my awareness of that clear in my first sentence, but I recognize the difficulties in communicating in a text-only medium such as this.

I guess I'm just saying "I agree"!

> The project is simply a website built with Flask, Apache, Redis, and TypeScript.

Clever.

Needs an RDBMS too, and also ELK stack for logging

> Needs an RDBMS too

You could add MySQL and you'd have a MYFART stack, which is pretty robust

Spring, hive, iis, and typescript makes for a more solid stack

Don't leave out enterprise distributed fp on jvm: clojure, rabbitmq, angular, postgresql.

Would be quite modern

I prefer the SQLite, Haskell, IIS, TempleOS stack

>TempleOS

Lol

Sounds like hot air to me

Yes.. but is it web scale?

You could save a lot of overhead using regular shell scripts (sh) instead of flask

why is apache needed? flask can directly serve the requests right?

I just can't get behind this stack, it stinks.

In other news, I'd love to include this as a stack proposal next time I'm proposing a project through our corporate sausage factory, just to see how far I can get it before someone realises.

> Apache sits in front of an application built in Flask which pulls its data down from Redis while rendering a template powered by TypeScript.

What an insane selection of components. I love it.

Sometimes, the name does all the heavy lifting. You just gotta figure out how to make the thing to fit the name.

This is actually a deep sentence.

I can probably find a job description calling for 12 years' experience with Fart Stack at this point.

I sprayed my coffee

K8s without yaml... Now that's done magic

The yaml exists but it isn't free.

Would like to meet the creator of this.

DevOps is not about Kubernetes seriously.

It's about process: testing, ci/cd process, monitoring, logging, error tracking,...

Very tangential, but I personally hated using Apache's WSGI module (or whatever it's called). Gunicorn was always just easier for me. Maybe it's just me, but I didn't like the mental overhead of Apache for anything other than a proxy, and at that point, nginx is my best friend.

Goofy article though - helps break up the day!

Not a huge fan of config for Apache, nginx or haproxy.

I always felt we needed something more ergonomic and production ready

Ahahahahahaha

Reply on news.ycombinator.com