Setting up ci/cd pipelines is like swimming through shit.

Even if you are successful, it still stinks!

I always love how microsoft azure plays in a part of ci/cd by creating a drag and drop job feature and configuration. You dont have to write a lot of yaml. But when it comes to writing yaml for Eks in Aws for github workflows i always pull my hair doing that.

you think so?

which technology are you using?

we're using gitlab, and CI is one of my favourite parts when setting up a new project.

-

especially since i remember the dark times before ci, where the build and deployment process was not just completely undocumented, but also required arcane environment setups nobody was able to reproduce. for a project we inherited, we had to keep an old workstation alive and untouched just so it was _possible_ to ship the product. and i mean "spining-rust-instead-of-SSD"-old.

Yep, it's shit.

But it does depend a great deal on exactly how shitty your organization makes it.

I use GitLab at home for my personal projects and even though one or two of them are orders of magnitude more complex than what I do on the job, the CI/CD setup for it is actually pretty terse, simple, and straightforward. I have no real complaints about it.

Contrast that with the utter garbage we have at work, where there are templates on top of templates that inherit this and integrate that and I'm looking at literally thousands of lines of YAML just to build what is, ultimately, a pretty bog standard Java webapp, and get it deployed.

I can do what it does by hand with maybe half a dozen commands at most.

In other words: they've taken what can be a fairly nice and elegant thing and turned it into a corporate nightmare of staggering proportions that I HATE dealing with every time I have to.

So, to a large degree, it's about how much enshittification has set in.

Hot shit with tetanus.

I love what they're able to do, but configuring other people's systems is not something I enjoy