This is a small experiment I’ve done to test out a system where we have multiple nodes and a single worker process, registered globally, and running a task every two seconds. The process can only be running in one of the nodes, so the task it runs doesn’t get duplicated in the other nodes. If one of the nodes dies, the remaining ones will try to restart the process.

Any feedback or suggestions will be greatly appreciated :)

https://github.com/bigardone/elixir-distributed-supervisors-test