feat: improved error messages + hedge racing + retry enhancements

[jorgecuesta]

This PR introduces several reliability and observability improvements to PATH:

1. Protocol Error Propagation

• Problem: When requests failed due to no available endpoints, users received a generic error: "protocol-level error: no endpoint responses received"
• Solution: Added SetProtocolError to the RequestQoSContext interface to propagate specific errors from the protocol layer to client responses
• Result: Users now see specific errors like:
• "no valid endpoints available for service: service X"
• "selected endpoint is not available: relay request will fail: service X endpoint pokt1..."

2. Hedge Racing (New Feature)

• Spawn a parallel "hedge" request after a configurable delay if the primary hasn't responded
• The first successful response wins; the other is canceled
• Configurable via retry_config.hedge_delay and retry_config.connect_timeout
• Track outcomes via X-Hedge-Result header

3. Retry Enhancements

• Time Budget: max_retry_latency skips retries when the failed request already took too long
• Endpoint Rotation: Each retry attempt uses a different endpoint
• Heuristic Detection: Retry on JSON-RPC errors hidden in HTTP 200 responses
• Observability: Track via X-Retry-Count and X-Suppliers-Tried headers

4. Heuristic Response Analysis

• Detect errors in response payloads despite HTTP 200 status
• Identify: JSON-RPC errors, HTML error pages, empty responses, malformed JSON
• Record correcting reputation signals for detected failures
• Support retry decisions based on payload analysis

5. Response Metadata Headers

Header	Description
X-Retry-Count	Number of retry attempts (0 = first attempt succeeded)
X-Suppliers-Tried	Comma-separated list of attempted supplier addresses
X-Hedge-Result	Hedge racing outcome: primary_only, primary_won, hedge_won, both_failed
X-App-Address	Application address used for the relay
X-Supplier-Address	Supplier address of the responding endpoint
X-Session-ID	Session ID for the relay

6. Health Check Improvements

• Add response heuristic analysis to health checks
• Record heuristic-detected errors to reputation service
• Better debug logging for health check validation

Configuration

services:                                                                                                                                                                       
- service_id: eth                                                                                                                                                             
  retry_config:                                                                                                                                                               
    enabled: true                                                                                                                                                             
    max_retries: 2                                                                                                                                                            
    hedge_delay: 500ms        # Spawn hedge after 500ms                                                                                                                       
    connect_timeout: 200ms    # Fast-fail on connection issues                                                                                                                
    max_retry_latency: 5s     # Skip retry if request took >5s                                                                                                                
    retry_on_5xx: true                                                                                                                                                        
    retry_on_timeout: true                                                                                                                                                    
    retry_on_connection: true                                                                                                                                                 

Test Plan

• Manual testing with ETH, Cosmos chains
• Batch requests verified working (IDs preserved)
• Protocol error messages propagate correctly
• Build passes
• E2E tests (some upstream endpoint flakiness noted)")