I love that I need to fucking babysit the fucking CI pipeline because, half of the times, it just RANDOMLY FAILS and I need to hit that fucking retry button like a 19th century kid respooling a loom.

why not just fix it 😐

Yes same here. Annoyance is beyond imaginable!

Same here.
Azure devops?

@Lensflare no. gitlab in public data center.

Don't pretty much all ci providers have an automatic retry option?

The amount of time our CI/CD pipeline has STOLEN from me is beyond reckoning at this point. There's ALWAYS some random failure or other to deal with, some service that isn't available, some resource constraint that sudden popped up, whatever. This whole "just click this one button and we're good" dream must be a reality SOMEWHERE, but I sure as hell have never seen it.

"You cannot disable the E2E tests".

Lol. You don't know my powers. XD
I disabled the E2E tests. Suddenly everything worked like a charm.

Small story of how I fixed a build pipeline for good. There was a lot of rage to deal with afterwards, but I was just too tired of that bukkake mess they called E2E pipeline...

(It was a selenium protractor angular thingamabob... If I had to do a metaphor, I'd describe it as an elderly dementia patient with a defilibrator suffering from dementia... If that thing worked it was a miracle).

Document failures.

If the same part fails, rip it out, let the madness begin and just say: This fucked up piece of garbage costs us money, either rewrite or it will stay off.

Custom build pipeline, custom versioning system. Component version is kept in two places, one of them gets updated one of them doesn't. DevOps team: "yeah we know it happens, its the developer jobs to make sure it was updated in both places".