chore: RPC nodes should detect network upgrade

[cloud-j-luna]

Currently, RPC nodes operate passively and may not be aware of impending or active network upgrades. The risk is that nodes may continue serving requests during network halt.

We need to design a mechanism for RPC nodes to autonomously detect when a network upgrade is active or imminent and transition into a well-defined “network upgrade detected” state. This way clients like console can implement appropriate measures to handle this, such as offline mode.

[cloud-j-luna]

As services start using the RPC proxy, one solution could be to implement this at the proxy layer which would be able to detect that a network upgrade is happening and either scale down or circuit-break until the upgrade is complete. This way services will get a corresponding status code and gracefully handle offline mode.