Fix critical RPC connection and performance issues

[amish777]

Overview

There's a critical issue with the RPC (Remote Procedure Call) functionality in the Terminal application that needs immediate attention. This is affecting users' ability to connect to various blockchain networks reliably.

Current Problems

• Intermittent connection failures when switching between networks
• Timeout errors during high transaction volume periods
• Failed state synchronization between client and node
• Inconsistent error reporting making debugging difficult for users
• Memory leaks observed during prolonged RPC connection sessions

Technical Details

1. Connection Management:

   • The current connection pool isn't properly handling network switching
   • Websocket connections aren't being terminated cleanly when idle
   • Reconnection logic has edge cases that cause indefinite reconnection loops

2. Request Handling:

   • Large batch requests are overwhelming some RPC endpoints
   • No proper throttling mechanism for request rates
   • Inefficient error handling leads to unnecessary retries

3. Performance Issues:

   • Memory usage grows over time leading to degraded performance
   • No prioritization for critical RPC calls during congestion
   • Lack of proper fallback mechanisms when primary RPC endpoints fail

Proposed Solutions

1. Implement a robust connection manager with:

   • Proper lifecycle management for RPC connections
   • Smart reconnection strategy with exponential backoff
   • Connection health monitoring and proactive failover

2. Enhance request handling with:

   • Request batching optimizations
   • Intelligent retry policies based on error types
   • Request prioritization system

3. Add performance improvements:

   • Memory usage optimizations
   • Connection pooling refinements
   • Implement multiple RPC provider fallback strategy

Acceptance Criteria

• Zero connection drops during network switching
• 99.9% RPC request success rate under normal conditions
• Graceful degradation under network stress
• Memory usage remains stable during prolonged usage
• Clear, actionable error messages for users when issues occur
• Comprehensive logging for debugging
• Load testing verification under high transaction volumes