feat: add state validation and metrics for replication

[bmuddha]

Problem:
Need to verify standby nodes are correctly replicating state and monitor replication health (lag, errors, throughput).

Proposal:
Implement blockhash verification at block boundaries, checksum computation for SuperBlocks, and expose metrics for replication lag, message throughput, and error rates.

Acceptance criteria:

• Blockhash verification against local observation
• AccountsDb checksum for SuperBlock comparison
• Metrics exported (lag, throughput, errors)
• Health check endpoint

Reference: MIMD-0019 Phase 4

Parent epic: #935
Depends on: #952, #953