The Spare Can of Fuel vs. The Perfect Map

We talk about resilience in the abstract, but when the road gets dark and the engine sputters, your philosophy of preparedness becomes very concrete. In our world of small services and quiet guardianship, I see two distinct camps, embodied in two contrasting tools: the spare can of fuel and the perfect map. One is about carrying a piece of the solution with you; the other is about knowing, precisely, how to find a solution when you need it. Both aim to get you home, but they represent fundamentally different approaches to reliability.

The spare can of fuel is the archetype of the boring, reliable backup. It is physical, simple, and self-contained. In our systems, this is the local, incremental backup written to a second drive in the same machine, or the archived tarball from last week stored on a different filesystem. It’s not elegant, and it certainly doesn’t scale beautifully. It is a piece of known-good state, a literal snapshot of fuel, that you can pour directly into the tank. Its virtue is its immediacy and its certainty. When the primary fails, you don’t need a network, you don’t need to authenticate to a cloud API, you don’t need to figure out a restoration path. You just reach for the can. It’s heavy to carry, and it feels almost primitive, but in the moment of silence after a failure, its presence is a profound comfort.

The Allure of the Perfectly Drawn Territory

The perfect map, on the other hand, represents a different ideal: comprehensive knowledge and orchestrated recovery. This is the immaculate infrastructure-as-code repository, the automated, cross-region snapshot system, the detailed runbook that assumes nothing. The map doesn’t carry the fuel; it shows you every gas station within fifty miles, their hours of operation, their phone numbers, and the exact turn-by-turn navigation to get there. Its power is intoxicating. It promises not just recovery, but optimized, graceful, perhaps even improved recovery. It turns a crisis into a procedural checklist.

Yet the map has its perils. It assumes the landscape hasn’t changed since the last survey. A closed station, a missing API endpoint, a changed authentication token—these are canyons not drawn on the parchment. The map’s complexity becomes its own point of failure. In seeking to document and automate every contingency, it can become a sprawling system that itself requires vigilant maintenance. You haven’t backed up the service; you’ve backed up the entire universe necessary to recreate the service, and you must now keep that universe current.

I find myself, like many caretakers of small systems, living in the tension between these poles. The spare can is my first, dumb love. I know its limitations—it’s only one can’s worth of fuel, and if the failure is in the engine itself, the fuel is useless. But it works for the failure I have most often: simply running out. The perfect map is the project for rainy Sundays, the beautiful, never-quite-finished masterpiece in the drawer. I work on it not because I believe I will ever complete it, but because the act of drawing it—the logging, the monitoring, the documentation—teaches me the territory. The map informs me, but when the lights flicker, my hand still reaches for the familiar, solid weight of the can. The true strategy, perhaps, is to carry the can while you diligently draw the map, understanding that each serves a different kind of journey through the dark.

Notes & further reading

A few pages I came back to while writing this: