scheduler service should work on cluster

At the moment, in a cluster, we can choose to persist the scheduler in a distributed DB (each node uses the same DB). This would probably throw an event on each node of the cluster depending on when it was persisted.

What we want is to be able to set a parameter (either in the contribution or pass it in the API) to:
1. throw the event on every node of the cluster
2. throw the event on one of the node of the cluster
3. throw the event on only the one that persisted the scheduler (or the one that has the contribution)

Apart from the case 3, the event should be thrown even if the server responsible for it shutdown.