The Bug I Fixed at 2 A.M. — Where Did the Credit Go?

The alert came in at 2:11 a.m. My phone buzzed twice on the nightstand. I grabbed it half-asleep — the database connection pool was maxed out, several core endpoints were already timing out, and the alert channel was filling up with red exclamation marks one after another.

A connection pool is basically the number of tables in a restaurant. Only so many guests can be seated at once. Once the tables are full, new guests wait at the door, and if they wait too long, they leave — that's the timeout. A full house isn't a small thing either: enough people waiting, and the front desk phone starts ringing off the hook, complaints follow.

I got up, opened my laptop, and started reading logs. About twenty minutes in, I found it — a background job wasn't releasing connections properly, piling up little by little until it filled every table. I fixed the config, killed the stuck connections, restarted the service, and watched the monitoring graph slowly come back down. Nearly two hours of it. By the time it was stable the sky was starting to lighten. I lay back down, barely closed my eyes before the alarm went off, and showed up to work the next day with the same dark circles as everyone who'd slept through the night.

At the morning standup, the project lead flipped through his slides and said, "There was a database issue last night, the team already handled it," then moved to the next slide. Nobody asked who handled it. Nobody asked how long it took or what got changed. I sat there and felt something drop — those two hours might as well have never happened.

I've turned this over in my head a lot since. Can't quite tell if it's resentment or just confusion. There's a default assumption in this line of work: do the job well and the result speaks for itself. The system's stable, the endpoints aren't timing out, the monitoring dashboard is all green — isn't that the hardest proof there is?

Eventually I understood: the result is proof, but proof doesn't open its mouth on its own. Someone has to say it out loud — what broke, how close it came to being worse, who fixed it, how much effort it took. I didn't say it, so the project lead had nothing to say either. All he had was that one line, "the team already handled it," to paper over the gap. He didn't steal my credit. Those two hours of information just sat in my head, and nobody took it — I simply never handed it out.

I've seen it work the other way. A colleague once cleared an incident and immediately dropped a message in the group chat: "Connection pool maxed out, endpoints were timing out. Root cause was a background job not releasing connections. Restarted the service, it's back. Adding a monitor for it." Three sentences, maybe forty seconds of typing. After that message went out, nobody needed to summarize what he'd done, because he'd already said it himself. Next time something similar came up, he was the first name anyone thought of. Come review season, other people have to try to recall what he actually did — every incident he'd handled that half-year was already sitting there in the chat log, easy to count.

This isn't just about whether you happen to mention it in chat. Whatever ends up on a resume, in a year-end review, or in a conversation about a raise, it's whatever got clearly said at the time. What never got said doesn't count as evidence later, no matter how technically hard it actually was — you can barely remember the details yourself, let alone expect anyone else to.

I still haven't fully kicked the habit. Last month there was another small incident. I fixed the config at 2 a.m., typed "handled" in the group chat, and rolled back over to sleep. The next day I remembered and went back to add a proper writeup — by then the incident was already ten-something hours old, the chat had moved on to other things, and nobody was going to click back and read it. Come year-end, trying to piece together what I actually did over six months, I had to dig through commit history myself just to remember. Nobody else stood a chance.

The smoother you get at the technical work, the more those extra few lines start to feel unnecessary, even a little like bragging. But the truth is, if you don't say it, the work might as well not have happened, as far as anyone else is concerned. Those two hours at 2 a.m. were real. Just not to anyone but me.

Next time I finish fixing something at 2 a.m. and I'm about to go back to sleep — I wonder if I'll remember to type those two extra lines first.