The Deploy That Ate My DoorDash Window

It was a Friday, which should have been a sign. The deploy was supposed to go out at four. A clean push, a green pipeline, a "have a good weekend" in Slack, and I'd be in the car by five with the app open and the dinner rush starting to stack up. I had it planned. I had the window.

At 5:47 p.m. I was still at my desk, still in the same hoodie I'd been in since nine that morning, reading a stack trace that was personally offending me. Something upstream had changed without a note in the changelog. The kind of thing where you stare at it long enough that the letters stop looking like letters.

The DoorDash window I'd been counting on — that specific two-hour block between dinner rush peak and the point where the orders thin out and the math stops working — was dissolving in real time, line by line, right there in my terminal.

Why I Was Counting on That Friday Window in the First Place

I've been doing DoorDash on and off long enough to know that not all hours are the same. Friday dinner rush in certain parts of Montgomery is reliable in a way that a Tuesday afternoon just isn't. I'd already mapped this particular window a few times. It wasn't a system. It was just a thing I'd noticed and started quietly building around.

The week had been expensive in the low-grade way weeks get expensive — a school thing, a car thing, nothing dramatic, just the usual slow leak. I wasn't in crisis. I just wanted to go out and make back something concrete before the weekend. That's the whole logic. It's never more complicated than that.

I'd even held off on eating dinner early so I wouldn't be making delivery runs on a full stomach going "man, that smells good" the whole time. That's the level of preparation I'd put into this. I was ready. The deploy was not.

What the Pipeline Did to My Evening, Specifically

The error was in an integration layer that another team owned. Technically not my problem. In practice, my code touched it, my name was in the recent commits, and my team lead — a perfectly reasonable person named Davis who I have no complaints about — sent me a message that said:

"Hey can you just take a quick look, shouldn't take long."

"Shouldn't take long" is a phrase that has never once been accurate in the history of software. I know this. I said sure anyway.

By six-fifteen I had confirmed the problem wasn't mine but had also, somehow, half-fixed it anyway because I was already in there and it bothered me. That's the thing about being wired the way I'm wired — I cannot leave a broken thing alone even when the broken thing has nothing to do with me and is actively costing me money to look at.

I got the deploy out at seven-oh-four. Davis said "legend." I closed the laptop and looked at the DoorDash app. The rush had peaked and was already sliding. I've written before about trying to make DoorDash make sense as a regular income source, and the honest answer is that it only kind of works when the timing is right. Seven p.m. on a Friday in my zone is not the same as five-thirty. The orders were lighter. The distances were longer.

I went out anyway. I don't know why I expected a different result. I think I just needed to leave the house.

I ran for about ninety minutes. Picked up four orders. One of them sent me to a part of town I don't love at night, and I sat in a parking lot for six minutes waiting for a gate code that turned out to be posted on the wall right next to the call box, which I would have seen immediately if I hadn't been staring at my phone. That's the kind of thing that doesn't show up in any math but costs you anyway. The hidden costs of a shift are always in the details you didn't think to track until you're already in them.

The Actual Numbers, Which I Am Posting Because That's the Deal

Four deliveries. Ninety-something minutes door to door. Thirty-seven dollars and change before you count gas.

I drive a sensible car and I don't live far from the zones I work, but gas is gas. I'm not going to pretend I did a precise per-mile calculation because I didn't — I never do, which is a problem I have acknowledged before and done nothing about. But I filled up two days later and the tank was lower than I expected it to be, and some portion of that is Friday night.

The day job cost me nothing in dollars. It cost me the window. Those are different kinds of cost and I am genuinely unsure which one bothers me more, which probably says something I don't want to examine too closely.

What I can say is that I've had better nights and I've had worse. The math on a DoorDash shift always looks thinner once you actually do it after the fact. Thirty-seven dollars is thirty-seven dollars. I'm not going to pretend otherwise, and I'm not going to tell you it was worth it, because I genuinely don't know.

What I Changed, and What I Definitely Did Not Change

I added a note to my whiteboard. It just says "FREEZE DEPLOYS AT 4." I don't have the authority to actually freeze deploys at four. It's not my call. But writing it down made me feel better, which is the whiteboard's entire job at this point.

The thing I didn't change is the habit of building around the day job's schedule like it's predictable. It isn't. It never has been. I have been doing this long enough to know that a Friday deploy is a gamble and I keep taking the gamble, which is either optimism or a learning disability and I'm not qualified to say which.

Davis, for his part, did not know he had eaten my evening. I didn't tell him. It's not his fault the pipeline broke, and "I was going to do DoorDash tonight" is not a sentence I'm going to say on a work call. There are things you keep separate. That's one of them.

The kids asked where I went when I finally got home. I said I had some errands. My oldest gave me a look that suggested she knew the errands involved a thermal bag, but she let it go. That was decent of her.

The deploy is in production. The integration layer is somebody else's problem again. I made thirty-seven dollars on a Friday night that was supposed to be better than that, and Monday came anyway, right on schedule, like it always does.

Note: Everyday Hustle Jeff is the written side of a digital series. These accounts are composite and dramatized. Any figures are what they were on the day, for one person, and are not a projection, a track record, or advice.