Home / Blog

Dead Man's Switch Pattern Explained

Originally a physical safety mechanism, in software a dead man's switch means: if I stop sending proof-of-life, assume something went wrong and act. For batch jobs, the “proof” is usually an HTTP ping at the end of a successful run.

Why it fits cron

Cron is pull-based and opaque. The switch inverts the model: your job pushes success outward. Missed pushes imply failure-by-silence, which is easier to reason about than parsing absence of log lines.

Implementation sketch

1. Define how often success should occur.
2. On success only, call a monitored URL.
3. Service alerts if the URL is not hit inside the window + grace.

Product angle: dead man's switch monitoring with CronCraft.

Not the same as uptime checks

Pinging a public website tells you the site responds. Pinging CronCraft tells you your batch finished — different signal, same transport.