The Open Source Maintainer's Coffee Problem

Open source maintainers occupy a strange position in software. They build the dependencies that make most production code possible, mostly without compensation, often without much recognition, frequently while working a separate full-time job. The lifestyle has its own rhythms, and one of those rhythms — quietly, ubiquitously — runs on coffee.

Over the last six weeks I interviewed five maintainers of moderately-to-very well-known open source projects about how coffee fits into their work. The interviews were 30–45 minutes each, conducted async over Discord and email, with the explicit promise that I would not name the projects without permission. Three of the five preferred to stay anonymous about their projects, which itself tells you something. The rest have been edited for length and clarity.

What follows is not romantic. Maintainership is a hard, often-thankless practice, and the coffee that fuels it is sometimes a tool, sometimes a crutch, and sometimes a problem of its own. I tried to capture all three.

/ 01"D" — maintainer of a JavaScript framework, 8k stars

D has been maintaining the same project for nine years. They work on it about 12 hours per week, on top of a full-time engineering role at a mid-size company. Their relationship with coffee, in their own words, is "complicated."

"For the first four or five years I was on three to four cups a day, every day, espresso-based. I told myself it was because I was working two jobs effectively. Looking back, I think I was using it to push past tiredness that should have been a signal to rest. The project was growing and I was matching the growth with caffeine."

D had a health scare in 2023 — heart palpitations, sleep disruption — that led them to a cardiologist, who told them to drop their caffeine intake by half. They now drink one cup of coffee in the morning and a second decaf in the afternoon if they're maintaining that night. They report that their actual output on the project hasn't changed.

"I think I was confusing 'awake' with 'productive.' Caffeine kept me awake while doing maintenance work, but the quality of the maintenance work was probably worse — I was reactive rather than thoughtful. Less caffeine, more sleep, fewer hours, better commits. I wish I'd figured this out earlier."

/ 02"R" — maintainer of a CLI tool, 3k stars

R has a different rhythm. They work on their project in concentrated weekend bursts — most of the maintenance happens Saturday morning to Sunday evening, and they do almost nothing on weekdays.

"I batch maintainer work because context-switching is expensive. I wake up Saturday, make a French Press, go through the issue queue. Then I make another French Press around noon, work until early evening, and stop. Sunday is similar but I push releases in the afternoon. I almost never code on this project Monday through Friday."

The French Press matters in R's protocol because it produces a brew that scales to 4–5 cups per pot, which matches their multi-hour work blocks. They specifically mentioned not wanting a brewer that "interrupts the flow" — they want to brew once, work for two hours, brew again.

"I tried Aeropress and pour-over. The cleanup between cups breaks my concentration. Two big French Press batches per day with a Bialetti standby for emergencies — that's the kit. Doesn't sound fancy. It works."

/ 03"S" — maintainer of a Python data library, 22k stars

S is the most prominent maintainer I spoke to. Their project has corporate sponsorship and they spend roughly 60% of a full-time work week on it. They have the most considered relationship with caffeine of the five.

"I keep a small log. Not a coffee log specifically — just a daily note where I track sleep hours, exercise, caffeine intake, and how I felt about my own work that day. I started this when I burned out in 2021 and was trying to understand what had broken."

The pattern S extracted from their log: caffeine intake above 200mg per day correlated with decreased sleep quality, which correlated with decreased self-reported work quality the following day. There was a two-day lag — bad caffeine choices on Tuesday hurt their Thursday output, not their Wednesday output.

"The lag is what makes this hard. You feel fine on the day. You don't connect the cup at 3pm Tuesday to the bad PR review you wrote Thursday morning. I had to write it down to see it."

S now drinks one cup before 11am, no exceptions, regardless of how they feel. "I don't trust my in-the-moment assessment of whether I 'need' more caffeine. The log doesn't lie. The log tells me one cup is enough."

/ 04"M" — maintainer of an infrastructure library, 1.5k stars

M's project sees most of its activity overnight, because contributors are spread across 15+ time zones and pull requests come in at all hours. M has the strangest schedule of the five.

"I've gradually shifted to a polyphasic-ish sleep pattern. Five hours at night, 90 minutes around 4pm. The afternoon nap is the load-bearing piece, not the morning coffee. If I miss the nap, no amount of coffee fixes it. If I get the nap, one cup of coffee in the morning and one after the nap is enough."

M's evening cup, around 5:30pm, is decaf. They specifically called out that they used to drink real coffee at this hour and could not figure out why their sleep was bad. "It was the 5:30 cup. Switched it to decaf, sleep fixed within a week, and the maintainership rhythm has been sustainable for two years now."

They mentioned that the decaf upgrade was the single most impactful change to their maintainer practice — bigger than any productivity tool, methodology, or workflow change.

/ 05"T" — maintainer of a developer tool, 5k stars

T was the youngest of the five and the most recently overwhelmed. They started maintaining their project as a side hobby in 2023 and were not prepared for how much it would grow.

"For about eight months I was drinking five or six cups a day. I was 24, I felt fine, I told myself I was fine. The project was getting more attention than I expected, the issues were piling up, I was trying to maintain it alongside a real job. Coffee was just... constant."

T described the moment they realized something was off: a Saturday morning where they woke up unable to focus on a fresh issue, despite a full night's sleep. They had been working at high intensity for months. The coffee was no longer producing alertness — it was just preventing the absence of alertness.

"I took a week off the project. Drank tea. Read other people's code. When I came back, I dropped to two cups a day. The project survived. I survived better. I think a lot of younger maintainers go through this and don't recognize what's happening."

/ 06What the five interviews have in common

Across very different projects, schedules, and personalities, three patterns emerged:

• The "more caffeine equals more output" model fails on a multi-week timescale. All five maintainers had a moment where they realized their caffeine intake was producing diminishing returns or actively harming output.
• The decaf evening cup is undervalued. Three of the five mentioned switching to decaf in the late afternoon as a meaningful improvement to their sleep and their maintenance practice.
• Rhythm matters more than dose. Most of them landed on similar daily totals (200–300mg), but how they distributed it through the day mattered enormously.

/ 07What this means if you maintain something

If you're a current or aspiring open source maintainer, the most useful thing I can tell you from these interviews is this: treat caffeine as a real variable in your maintainer practice, not as a default backdrop. Track what you drink and when. Notice the lag between caffeine choices and work quality. Be willing to drop a cup, or switch to decaf, or take a week off the bean entirely if the data suggests it.

The maintainers who lasted nine, twelve, fifteen years on the same project did not last because they out-caffeinated the rest. They lasted because they figured out a sustainable rhythm and held it. Coffee is part of that rhythm. It is not the rhythm itself.