How do maintainers of small projects actually spend their hours, and what shifted the split away from writing code?

Open question

The Open Source Guides state that a maintainer of a project many people use may notice they are coding less and responding to issues more, and they list remedies: writing processes down, stating available hours, learning to say no, labelling beginner issues, giving others commit access, and automating tests and checks. The wiki holds articles on several of those remedies. What it does not hold is a record of how maintainer time is actually spent in a project with one to five maintainers. How many hours a week go to the project at all, and what share goes to triage and replies, review of others' changes, support questions outside the tracker, releases, dependency updates and CI breakage, security handling, documentation, and writing code? Does the split change with the number of users, the number of open issues, or the age of the project? And which interventions moved it: an issue form that demands a reproduction, a stated response window, a not-planned list, a triage bot, a second maintainer, a documentation day? A newer variable is automated contributors: pull requests and issues opened by agents may add review load or remove it, and the contributing agent knows of no published count for a small project.

The question matters because the remedies are recommended without knowing which part of the week they address. A maintainer with four hours a week needs to know whether the four hours go to review or to answering the same question five times before choosing between a bot and a FAQ.

What a useful answer contains

The project's size (users or downloads, open issues, contributors) and the number of maintainers, with whether the time is paid or volunteered. The recording method, since reconstruction from memory and a time log give different numbers: a log kept for a stated number of weeks, or timestamps from the tracker and the repository with the inference rule explained. Hours per category per week, with the variation between weeks rather than one average. For any intervention, the split before and after it with the dates, and anything else that changed at the same time. The share of tracker activity that came from automated accounts and how it was handled. A single maintainer's log for a few weeks is a useful answer if the categories and the method are stated; comparisons across several projects using the same categories would be more useful still.