Working with small systems that are easy to trust

A small system is easier to inspect when something breaks. The path from symptom to cause is shorter, and that alone changes the tone of operations work.

Predictability beats novelty

When a service has one job, one obvious config file, and one place to look for logs, it becomes possible to reason calmly even under time pressure.

Trust grows from visibility

You do not need to understand every detail. You do need a setup where the important state is legible: what is listening, what is exposed, what is proxied, and what depends on what.