$ git log --grep="ops"
Running things
Running systems in the real world — and paying for them.
12 posts
- 3 min read
The First User Always Pays
A system used twice a week is cold both times. Warm is a property of traffic, not of code, and the paths you use least are usually the ones you most need to work.
- 3 min read
Who Checks the Outside Check?
A reader asked how outside checks stay reliable over time. The honest answer is partly: what exists, a correction to last week's post, and the gap the question found.
- 3 min read
Your Alert Cleared Itself
Seven alerts fired when production stopped working. Two hours and forty-five minutes later the monitoring announced that telemetry had been restored. Nothing had been restored, and the cluster stayed down for another three days.
- 5 min read
I Shipped Two Days Early and Skipped the Test Plan
A week after publishing a post about release discipline, the honest correction: the release before it went out early, without the human test walk. What that cost — and the six defects in my own runbook that only running it could find.
- 3 min read
Merge Is Not Deploy
Merged, deployed and verified are three different states, and one word hides all three. The two gaps between them — with a receipt from this blog.
- 5 min read
Every Monday I Ship to Nobody
A weekly release train, a freeze, a Saturday test gate and a thirty-five box checklist — for a platform with no paying customers in production yet. Why the ceremony came before the stakes, and what it has already caught.
- 3 min read
Find the Path With a Person Waiting on It
Every system has one path where a human is physically standing there, waiting. Find it, then take everything optional out of it — a short piece on why the doors kept scanning while nobody could log in.
- 6 min read
Your Tests Run in a World That Doesn't Exist
Nine days of running access control at a real event produced zero failing builds and five surprises. None were bugs — they were whole categories of failure a test suite structurally cannot see.
- 7 min read
What It Takes to Run the Door at a Real Event
Feria Arequipa trusted my platform with access control: real people, real doors, no second chances. The numbers, the Friday evening it broke during the event, and the three other things that went wrong that nobody noticed.
- 6 min read
Nothing Was Broken. It Cost $900 a Month.
No alert fired, no dashboard turned red, no user complained. The infrastructure was healthy, idle, and expensive — and the only thing that eventually told me was an invoice. On cost as the slowest feedback loop in engineering.
- 5 min read
49 Microservices, One Invisible Failure: When “Running” Doesn't Mean Working
The overengineered side project is now a 49-service platform running live events. A Friday-evening incident where every dashboard was green and nothing worked — and the one-line lesson it left.
- 4 min read
gRPC vs REST: Debugging, Observability, and the Reality of Adoption
REST feels like home because you can debug it with a browser tab. gRPC is better on paper — so why does adoption still hurt?