Progol is a library that enables processes to find and communicate with each other over a network.

Process discovery (which are the processes that should be in this network?) and validation (which of those endpoints are currently available?) are two distinct stages.

At the lowest level, a Discovery component determines the set of ideal peers from some communications substrate. For example, a static list of endpoints, or by reading messages of a specific format from a multicast group. A couple of Discovery mechanisms are included, but you're free to write your own.

At the next level, a Validation component receives the set of ideal peers from the Discovery component, and "pings" each one to determine if it's alive. Progol defines a ping as an HTTP GET on the process endpoint at ValidationPath, and success as an HTTP 200 OK response.

At the top level, your application code should subscribe to updates from the Validation component, and make whatever decisions are appropriate for the current state of the network. Deeper introspection of peers is up to the application to define and verify.

Progol enforces no architectural restrictions on process discovery. The Multicast Discovery component that ships with Progol is an example of P2P discovery. It's also possible to write a Discovery component that appeals to a defined master coördinator.