Nobody Showed Up to Practice (Again)

The Slack message came in at 4:47 p.m. on a Thursday, which is the worst possible time for anything to come in, because it means you’re already mentally in Friday, already thinking about the thing you’re supposed to do after work, and now that thing is not going to happen.

“Production’s down,” Marcus wrote. “Payment processing endpoint returning 500s. Client is aware.”

I stared at my phone. The Merge Conflicts had rehearsal at 8. We’d cancelled the last three. Actually, not “cancelled”—those were on me. Last month I’d bailed twice because of “urgent meetings,” which is a phrase that means nothing and everything. Three weeks ago I’d texted the group chat at 7:55 that I wasn’t going to make it, and nobody even responded because they’d already stopped believing I would show up.

This time was different. This time it was real.

I got up from my desk and walked to the dev area. Dex was already deep in something, his monitor throwing light across his face in that way that means he’s been staring at code for too long. Marcus looked up with the expression he gets during incidents—hunted, like he’d personally written the bug.

(There should be a support group. “Hi, I’m Paul, and I have, on at least four separate occasions this year, stayed an extra ninety minutes ‘just in case’ a fixed problem un-fixes itself. It has never once un-fixed itself. I keep showing up anyway, like a man checking whether a door he already locked is still locked.”)

“What’s the scope?” I asked.

“Not sure yet,” Dex said without turning around. “Marcus is checking the logs. Could be the payment gateway integration we pushed Tuesday, could be something in the auth layer.”

I nodded. I already knew what I was going to do. I was going to call the client—Vantage’s biggest contract, the one that pays for like thirty percent of the office—and tell them we had eyes on it. That part was my job. That part made sense.

I pulled up the client call and got Patricia from their ops team on the phone within two minutes. She was already in a bad mood, which is what happens when a payment system goes down: all the bad moods in the world converge at one point.

“Paul, it’s been down for nine minutes,” she said. Not angry yet. Worried. That’s actually worse.

“I know. I’ve got our senior engineer on it right now. First thing we need to know is what changed. I’m going to stay on this with you until we’ve got it back up.”

That was true. I would stay. But the way I said it—”I’m going to stay on this with you”—made it sound like I was going to debug it, like I was the one who was going to fix it. I’m not. I haven’t written a line of production code in four years. What I was actually going to do was sit in this chair and exist, occasionally asking Dex questions he was already thinking about, and relay information back to Patricia that Dex had already figured out.

By 5:30, Dex had narrowed it down to the payment gateway integration. By 5:45, he’d found a caching layer that wasn’t invalidating properly when a new API key rotated through. By 6:10, he had a fix ready. By 6:20, we’d pushed it and the endpoint was returning 200s again.

I told Patricia it was fixed. She thanked me like I’d personally written the code, which I didn’t, and it felt good and bad at the same time.

“I’m going to stick around for another hour,” I told Dex and Marcus. “Just in case there’s a cascade or something rebounds.”

Dex nodded. Marcus said, “Yeah, good call,” with the tone of someone who really does want to leave but respects the cautious choice.

I sat. I refreshed the monitoring dashboard. Nothing broke. I refreshed it again. Still nothing. By 7:15, the incident was genuinely over. The system was stable. There was no technical reason I needed to be there anymore. Dex could spot a problem if one appeared, and even if one did, it was his code, so it made sense for him to fix it.

But I stayed.

I refreshed the dashboard. I Slacked Patricia a “still monitoring” message at 7:30. She appreciated it. At 7:45 I finally texted the band group chat.

Kyle actually has a day job and legitimate excuses. That’s what made “next time” land the way it did — not as a complaint, just as a quiet tally mark on a scoreboard nobody else was keeping.

By 8:15, when I finally actually left, I realized something that sits with me as I’m writing this: the incident needed me for forty-five minutes. Forty-five minutes to validate the client, field their panic, make the decision to bring Dex in, and keep them updated on the scope. After that, it needed Dex. It did not, at any point, need me refreshing a dashboard and existing near Dex while he worked on something he knew how to do better than anyone in the room.

The other two hours and fifteen minutes I spent hovering were just habit. Just the thing I do because I’ve done it before, because it feels like I’m being responsible, because it’s easier than saying “I’m not actually needed here anymore” and leaving.

Easier than admitting I’d cancelled the band because I wanted to, not because I had to.

“Easier than admitting I’d cancelled the band because I wanted to, not because I had to.”

My mistake wasn’t staying during the incident. My mistake was staying after the incident was fixed. And the thing that got me to stay wasn’t diligence—it was the fact that I’d already told myself I was cancelling the band, so why leave? The incident was the excuse, not the reason. I would have found another excuse if the incident hadn’t happened. I’ve been doing exactly that for four months.

The pattern is this: a real emergency happens (and they do happen, that part was genuine), but then I mistake “being in the room while it gets fixed” for “being necessary.” So I stay, because staying feels like I’m doing something, and more importantly, because it lets me tell myself that I’m not abandoning my responsibilities. Except I am abandoning something. I’m just doing it to a band that doesn’t have a client account manager to escalate to when I ghost them.

If you freelance, you’ve had my exact conversation, just with a client instead of a VP. An emergency happens—a bug goes live, a deadline shifts, something breaks. You drop everything. That part is real and usually correct. But then the emergency gets handled, and you stay online “just in case,” refreshing the dashboard, staying in the email thread, telling yourself you’re still needed. Meanwhile, the thing you cancelled—the other client project, the proposal you were going to write, the thing that’s actually your business and not just your reaction to someone else’s crisis—sits there waiting.

So here’s what I’m doing this week, and what you should do too: the next time an emergency happens, ask yourself specifically—not vaguely, specifically—”what do I personally need to be doing right now?” Write it down if you have to. Once that thing is done, ask it again. And when the answer is “nothing, I’m just watching,” leave. Don’t stay. Don’t refresh the dashboard. Go do the thing you cancelled. The incident doesn’t care if you’re watching it resolve. The incident only cares if you’re the only person who can fix it, and once you’re not, you’re just lying to yourself about why you’re still there.