The first useful health check is external.
A daemon can be alive while the service is unreachable. Start with the interface people actually depend on, then work inward.
Read field note →Field Notes are deliberately small. They record an observation, a question or an operating pattern without pretending that every problem requires a framework.
A daemon can be alive while the service is unreachable. Start with the interface people actually depend on, then work inward.
Read field note →Separate DNS, connection setup, transport and processing instead of hiding them behind one number.
Read field note →The worst time to discover that a backup is incomplete is during the incident it was supposed to solve.
Read field note →Absence of errors can mean stability, missing instrumentation, or simply that no useful request reached the system.