Blog Archive
People Are Not Cattle
The 'replaceable cog' theory of management is the cattle model applied to humans — interchangeable, nameless, swap any one for another. It's true exactly as far as the work is a stateless, commodity replica. Your best people are not. They carry state you cannot reproduce — context, judgment, relationships, the undocumented why — which makes them pets, not cattle. Treat a pet like cattle and two things happen: you discover the replacement was never interchangeable, and, because this particular pet is self-aware and can quit, it notices first. And the cruel twist: the commodity work is exactly the work automation takes.
Read Post
Infrastructure Is Not an Application
We took the software-engineering playbook — DRY, long-lived feature branches, package managers, dynamic in-cluster orchestration — and pointed it wholesale at running infrastructure. It keeps breaking, and not because the tools are bad. An application is a stateless artifact you stamp out in identical, disposable copies; infrastructure is stateful, singular, and alive in production. The patterns that make app development fast make operations fragile — and knowing which is which is most of the job.
Read Post
You're Not Writing Code. You're Running a System.
The most important change an engineer can make isn't a new language or framework — it's a single sentence about what they think they're doing. 'I am writing code' makes the diff the deliverable and production someone else's problem. 'I am running a system' makes the live behavior the deliverable and the merge the starting line. Same keyboard, same person, a completely different job — and nearly everything good downstream flows from the switch.
Read Post
Maintainability Against the Unknown Unknowns
You can harden a system against the failures you can imagine. You cannot harden it against the ones you can't — and those are the ones that take you down. The only defense against a failure you never predicted is a system you can understand, change, and recover fast enough to survive the surprise. That capacity has a name: maintainability. It's the most undervalued property in engineering, because its payoff is invisible until the day it's everything.
Read Post
Manual, Automated, Autonomous
There are three rungs on the operations ladder — do it by hand, run a script that does it, or build a system that keeps it done without you. Most teams climb to the middle rung, call it 'automated,' and stop — never realizing the rung above is where reliability actually lives. The Google SRE book laid this out a decade ago. It's the most-cited, least-applied book in the field.
Read Post
The Balanced Org
Separate offense (product, engineering, sales) from defense (infrastructure, IT, security, compliance), hold them as peers in dynamic tension under the CEO, and put the right temperament in every seat. A visual essay on org design — and why no single 'balanced' leader can hold both halves of the job at once.
Read Post
The Digital Plumber
Ask an SRE what they do and you'll get a hundred answers. Here's mine: we wear three hats — advisor, plumber, and firefighter — and not one of them is 'product.' We are the people who make sure the product can be built cheaply, shipped reliably, and kept alive at 3 a.m., whatever the product happens to be. Because after enough years, you stop seeing products at all. You just see systems, and bytes moving through them.
Read Post
Chasing Nines
Your availability target — how many nines you promise — isn't a line in a contract, it's an architecture decision. Get it wrong in one direction and you fail the SLA you sold. Get it wrong in the other and you burn money buying nines you never promised. I once looked at a stack paying $10,000 a month for high availability it did not actually have.
Read Post