feature: rpc proxy/caching layer

[nick1udwig]

Is your feature request related to a problem? Please describe.
We are quite reliant on ETH RPC & hit it a lot. This leads to some fragility & dev annoyances. It is also hard to debug because we have many nodes per API key.

Describe the solution you'd like
There are at least two main things we'd like.

1. Some type of proxy system where each node can have its own unique identifier for the RPC calls. In this way, if a specific node is in a bad state, we can determine which node is the bad one.
2. Some type of caching system. Each node needs to hear about the same events. If we could cache these, it would greatly reduce the load on the actual RPC.

Should this live inside or outside of Kinode? Are there specific pieces that should live inside and others outside?

FYI @dolled-possum @bitful-pannul @dr-frmr