What Twin Toddlers Taught Me About Incident Response

I have spent years in tech learning how to handle production incidents. I have run postmortems, written runbooks, and sat through at least one meeting titled "Blameless Retrospective" that was extremely not blameless. None of it prepared me for the real training ground: two toddlers who can break two different things at the same time.

Here is what twin toddlers taught me about incident response.

1. Severity is not about volume

Every on-call engineer learns that the loudest alert is not always the worst one. Same with twins. One twin screaming in the living room is usually a Sev 3: a toy was taken, justice must be served, and everyone will be fine in ninety seconds.

The silence from the other room? That's the Sev 1. Silence means a marker, a tube of toothpaste, or a determined little scientist running an experiment on the couch. I've learned to treat quiet like a pager going off with no message body. Drop everything. Go look.

2. You will always have two incidents at once

Nobody warns you about this. In a normal workplace, incidents arrive in an orderly queue. With twins, they are concurrent. One needs a snack right now, the other has discovered gravity applies to yogurt, and you are standing in the middle holding exactly one paper towel.

This is where triage matters. Who is in danger? Who is merely sticky? Who can wait ninety seconds with a cracker as a temporary workaround? Yes, the cracker is a hotfix. No, it is not the permanent solution. We'll refactor at dinner.

3. Don't fix what you haven't diagnosed

Early on, I'd rush in and "fix" things before understanding them. Offer a snack when the problem was tired. Offer a nap when the problem was hungry. Offer a hug when the problem was that I had cut the sandwich the wrong way, which, for the record, is a real and valid category of outage.

Now I ask the same questions I'd ask on a bad deploy. What changed? When did it start? Is this one twin or both? Did anyone recently get a different-colored cup? Most issues are a config change nobody told me about.

4. Communicate, even when you have nothing

The best thing I learned from real incidents is that people can handle bad news; they can't handle silence. The same applies at home. When I'm wrangling one twin and the other is waiting, I narrate: "I see you. I'm coming. You're next." That one status update buys me a surprising amount of goodwill.

At work I do the same thing, just with fewer capital letters: "We're aware, we're investigating, next update in 30 minutes."

5. The postmortem is where you actually get better

After a rough evening, my partner in all of this and I will sometimes do a tiny, tired, informal postmortem over leftover pasta. What went wrong? What went right? What do we do differently tomorrow?

• Root cause: nap skipped.
• Contributing factor: grocery store at 4 p.m.
• Action item: never go to the grocery store at 4 p.m.

We will absolutely go to the grocery store at 4 p.m. again. Action items are aspirational. That's true in tech too, and we all know it.

6. No heroes, just a team

The thing I miss most in bad tech incidents, and the thing I'm proudest of at home, is the absence of heroics. The best teams don't have one person who saves the day at 3 a.m. They have handoffs, rest, and someone who says "I've got this one, go sit down."

Parenting twins is a two-person rotation with no time off, and I've never been more grateful for a good handoff. "Tag, you're up" is the most romantic sentence in my house.

The uptime I care about

I used to measure success in nines: 99.9% this, 99.99% that. With twins, I'll happily take a different metric. Did everybody get fed? Did everybody get hugged? Did anyone laugh so hard they fell over?

Some days the answer is yes to all three, and the dashboard is gloriously green. Other days it's a sea of red and we just try to get everyone to bed in one piece. Either way, tomorrow there's a new on-call shift, and honestly? I wouldn't give up the pager for anything.