Issue Description

While working with WebSocket and WebRTC transports, I noticed several reliability issues in connection establishment that could cause failures in production environments.

Problems Fixed

1. WebSocket Issues:

• No retry mechanism for failed connections
• Basic error handling without proper cleanup
• Missing connection timeouts

2. WebRTC Issues:

• ICE candidate gathering could hang indefinitely
• No proper timeout for SDP answer
• Connection state transitions weren't properly monitored

Solution

Added robust connection handling:

• WebSocket: Added retry mechanism, proper timeouts, and better error handling
• WebRTC: Added ICE gathering timeout (30s), connection timeout (60s), and proper state monitoring

Testing

Tested with:

• Slow network conditions
• Network interruptions
• Various failure scenarios (ICE failures, timeout scenarios)

The changes significantly improve connection reliability in unstable network conditions.