This topic was already discussed here a while back but up until now I haven’t found a viable solution to the problem:
We would like to host 2 (or more) instances of a service that reads messages from a stream and then performs some work on them. Ideally the instances would share the load of incoming messages and if one goes down, the rest would take over. Messages should only be handled once. As Event Store currently doesn’t care how many subscribers are reading from a stream, the responsibility of checkpointing and failover handling falls to us. I think Greg mentioned a while back that there was some work under way to better support such a scenario with ES, so my first question is does anybody know the status on this is?
Did anybody find an easy solution that doesn’t involve implementing your own clustering logic like heartbeat messages, finding a quorum, etc.