Add multi-tier RPC failover: primary, secondary, fallback polling per chain #35
Description
Summary
Generalize the RPC failover logic to support primary + secondary providers per chain, with automatic failover and fallback to HTTP polling.
Changes
- Per-chain RPC configuration: primary URL, secondary URL, poll interval
- Shared health-check logic (reuse from Base implementation)
- Failover: primary to secondary to polling
- Metrics: failover count, current provider, health status per chain
Tests
- Primary failure triggers failover to secondary
- Both providers fail triggers polling mode
- Recovery restores WebSocket mode
Acceptance Criteria
All chains have resilient RPC connectivity with automatic failover.