The deployment pipeline crashed again

>the deployment pipeline crashed again
>spend 20 hours per week maintaining the """automated""" deployment pipeline

previously we just copy pasted zip files and everything just worked

Attached: 1643812211184.png (3000x1950, 164.71K)

getting paid to do nothing and you complaining?

dumb nigga

how is doing actual work 20 hours per week nothing?

Fix the damn deployments devops janny so it actually becomes automated then.

How is your pipeline constantly crashing though? Is the CI blowing up? The CD? If you had a process that just worked before then why didn't you just script said process and run it through one of the CD programs?

you need to subscribe to a service and pay for an enterprise account that uses GitHub in your pipeline. It should cut your down time from 20 hours per week to just 5 or so for a million dollars per year.

"devops" was a mistake

This

>he fell for the devops janny helltrap
Retard.

write application with:
>compile time safety
>integration and unit tests with 100% coverage
>flexible, extensible, modularised

write deployment pipeline with:
>4000 lines of .yaml
>untestable
>half of it copy pasted
>half of it duplicated
>poorly understood dependencies
>poorly understood environment
>no awareness of security risks
>feedback cycle of 30 minutes
>poor managemt of secrets
>pipeline could break at any time due to external dependencies and idempotent builds
>nobody can touch it ever again in-case it breaks and takes weeks to fix
>spend more time fixing the devops pipelines than the application itself

"people" prefer their """automated pipeline""" over ssh or ftp because it lets them spend 10x the money on web services for infinite scale they'll never need and wouldn't work anyway because of other bottlenecks they haven't considered

Attached: 914-9144208_ck-food-cooking-png-wojak-fat-crying-crying.jpg (820x822, 195.35K)

Yep, pretty much this.

Is it really?

I have an interview with IBM for a DevOps position. Should I consider other positions?

Then make it so it doesn't crash.
> I write code and it's shiiittt
> it crashes all the time but it's not my fault!
> users dare to use it wrong!!!!!!
The only cure for programmers this bad is firings.

>IBM
>DevOps
I am 100% positive that IBM is only good for comfy committing 2 lines of code per week while bullshitting everyone on dailies how you look into issues. Am I wrong? I had an experience like this in an old telecom provider.

IBM is actually a super innovative company. Its just credential walled.

Super innovative company where the average age of employees is 50+?

IBM is literally the only company spearheading quantum computing. The underlying scientific libraries of Qiskit are written in C++ too

this. are you incompetent? I have several working never-broken CI/CD pipelines and I'm not even devops

>the only company
Haven't heard of IBM shipping commercial quantum PCs. Maybe in 2018? Can't remember, even.

Why would your deployment pipeline be untestable? And if you have a single .yaml file that's that long or constantly copy pasting shit then it sounds like you're not taking advantage of whatever your CD gives you for reusing code.

For example if you're using Ansible you could just start defining your own modules for repeated pieces of code. I think Kubernetes has stuff like Helm, and I'm sure terraform has their own stuff for code reuse.

As for testing your deployment pipeline, once you break it down to separate pieces you can test each of them. If you need to make sure a server was started, then make sure it's running by checking ps or something that the server's up and running. You don't need something fancy, it's probably good enough to just run ps | grep "$YOUR_SERVER_PROGRAM_NAME" or just run nc to whatever port you run your server on.

It just sounds like you've created a monolithic beast in your deployment code rather than your application. Just relax and start refactoring it and add tests like how you would for an application.

>application
It would be good on a resume. Depends on how experienced you are and what your other options are. I landed a comfy job with a midwest giant landlord. Jews collect giga-tons of money and pay me 6 figures to drip out features to an outdated system.

I’m a recent grad. Only experience was at an internship for computer vision related problems and research on probabilistic Boolean networks and deep learning for auto regulating networks.

I’ve honestly forgotten a lot of stuff already studying for LSAT. Got a 162 on a mock test. But I’m kinda broke so I applied for a couple of tech jobs.

>run CI in sourcehut builds
>can SSH into the machine and inspect/fix if something goes wrong
FAGMANcucks BTFO

truest thing ive seen on Any Forums