Heartbeat

A distributed system runs on many machines. Any of them can stop working at any moment. The other machines have to notice, and they have to notice soon.

Noticing is harder than it sounds. When a machine stops answering, you do not learn why. It may have crashed. It may be alive but very slow. The network path to it may be broken while the machine itself is fine. From the outside, all three look the same. You get silence.

A heartbeat is the usual way to deal with this. It is a small message that a machine sends on a timer to say "I am still running".

The message is deliberately small