Support for worker based deployment

[mgoerens]

[genofire]

With ReplicaCount or HPA?

[mgoerens]

I haven't put much thoughts into it yet.

First step for me is to read about how to setup workers for Synapse. At the moment I'm aware of the existence of this setup, but I need to learn the details (and run it) before deciding how to implement it here.

Reg. HPA vs ReplicaCount: I think it would make sense to set a default replica count for the Deployment/StatefulSet (TBD) and have a HPA scaling up or down depending on the load / network traffic (pr other custom metric, TBD) of the worker Pods.