I'm having similar issues, and i think i found a way out: stop playing their game. Don't make enterprise software. Don't write unit tests. Don't accept pull requests. Simply write software for yourself and have fun doing it. Forget refactoring code into modules, just fucking code. Don't worry about deployment with k8s, just copy the Python script to your production folder and run it. Fuck all that shit about git bran…
This is a great point, laced with justified indignation. As a graybeard, I've seen multiple generations of "how to do software" and the most recent are the least fun. A lot of this is driven by the agile approach. It's tailor-made for dev burnout: from the endless tight cycles that force people into an infinite loop of productivity with scant satisfaction that comes with "completion", to all of the tools and philosop…
What I really hate is filling in form A, which directs you to form B, which you fill in to see whether form C is required. Form B was put in place by a legal team, who don't provide any point of contact, and who are not your friendly local legal team.
Because the process of filling out forms is so time-consuming, your engineering team uses Asana to track it. However, your PMs use an excel spreadsheet, the legal team uses JIRA, your copywriters use a google doc, and the teams that own forms B and C use separate internally created tools. You update your form-filling progress in 6 different places, some of which have bugs and others of which aren't actively monitored, so you also have to use email/slack/skype messages to follow up the right people. Some of the right people are actually the wrong people. Some of the right people reject your proposal because they didn't read it properly. A few of them reject your proposal for reasons that are actually valid, but which you could not possibly have known, because they're based on tribal knowledge which is not documented anywhere.
Filling in the forms and fixing the issues takes literally 3 months and at least 4 group meetings as well as several skype/zoom calls. One day you are finally allowed to write the code. It's 150 lines including tests, across two services, and you're done in two days. Everything that caused your proposal to be rejected would have been found in development. You quietly wonder whether you would have been happier as a bricklayer.