I've seen a claim that developers waste about 75% of their time maintaining their development environment rather than actually writing code. I'm curious about how this plays out for teams in real life. What are typical time sinks that you experience? For example, things like Docker environments breaking unexpectedly, dealing with dependency updates due to security alerts, debugging CI/CD pipelines, or onboarding new developers to local setups.
For those of you in established codebases, how do your percentages break down for your week?
- How much time do you spend on pure feature development?
- How much time goes into maintenance of your tools and environment?
- At what point do you think it makes sense to completely rebuild your setup?
Also, do you think the fragility of environment configurations is just a reality we face, or is there a documentation issue that can be fixed?
5 Answers
While the 75% claim seems a bit exaggerated, it's true that established teams often have that feeling. My experience is more like this: 40-60% of my week is spent coding, 20-30% on environment issues like Docker and CI/CD frustrations, and 10-20% is taken up by meetings and context-switching. Most onboarding pains come from existing debt rather than inherent fragility, and I think it makes sense to consider a rebuild when onboarding takes more than a day or when no one fully grasps the pipeline.
Yeah, that's part of it. If you're not keeping track of multiple changes, things can fall apart quickly. Also, the suggestion about a dedicated maintainer plus automation seems spot-on!
I think the key lies in how organized your environment is. With proper configurations and good practices, I usually spend only about 5% of my time managing the environment as long as it’s stable. If it needs constant tweaking, then that’s a sign there’s a deeper issue with the setup.
It really varies based on your environment. In my last job, I was split evenly between coding and handling server updates, git issues, and our never-ending ticketing problems. But at my current gig, I find myself mostly fighting with PMs about the details of change requests. I barely have to mess with the setup unless I need to restart my Docker containers.
To be honest, if we're just talking about development work, I dedicate 100% of my time to pure coding. My team has a separate group for managing environments and deployments, so I just focus on my code and what I need for testing.
Definitely, but I think your local database setup still counts as part of your dev environment. It impacts your workflow regardless!
It really depends on the structure of your team and the complexity of the project. In a lot of places I've worked in, that 75% figure feels accurate. Developers often avoid fixing technical debt because it's tough to justify that time to management. But this is why roles like DevOps are so crucial—they help streamline those processes so we spend more time building rather than troubleshooting.
I get your point, but can you clarify what you meant about rot vs. fragility? Are you suggesting that too many PRs merging at once causes bugs related to contextual loss?