Optimize getting contract state and node state

[volovyks]

Description

Ideally, such operations should be placed in a separate thread that is not blocking protocols.
It has not been tested, but it can have a stronger effect when 1+ node is offline since requests are taking more time.

[volovyks]

We are not doing all /state calls in parallel. So we can optimize it significantly.

[volovyks]

Dashboard idea: once this is done, add metrics about elapsed time since the last contract and node state update. And the alerts on top of it. That will allow us to track RPC and other issues more easily.