improvements mid apr 2025

[lukaszraczylo]

GraphQL Monitoring Proxy - April 2025 Improvements

This document summarizes the changes in the improvements-mid-apr-2025 branch compared to the main branch.

Circuit Breaker Implementation

A new circuit breaker feature has been added for fault tolerance using the Sony GoBreaker library:

• Circuit breaker trips after configurable consecutive failures and automatically recovers
• Includes fallback mechanism to serve cached responses when circuit is open
• Configurable via environment variables:
  • ENABLE_CIRCUIT_BREAKER
  • CIRCUIT_MAX_FAILURES - Default: 5
  • CIRCUIT_TIMEOUT_SECONDS - Default: 30
  • CIRCUIT_MAX_HALF_OPEN_REQUESTS - Default: 2
  • CIRCUIT_RETURN_CACHED_ON_OPEN - Default: true
  • CIRCUIT_TRIP_ON_TIMEOUTS - Default: true
  • CIRCUIT_TRIP_ON_5XX - Default: true
• Full test suite included for circuit breaker functionality

Memory Cache Improvements

Enhanced memory-aware caching with:

• Configurable memory limits through:
  • CACHE_MAX_MEMORY_SIZE - Default: 100MB
  • CACHE_MAX_ENTRIES - Default: 10,000
• Memory monitoring for cache utilization tracking
• Smart eviction strategies to prevent out-of-memory situations
• Enhanced memory usage tracking and metrics reporting
• Compression for large cache entries
• Tests for cache eviction and compression strategies

GraphQL Parsing Optimizations

Performance improvements in GraphQL handling:

• Query parsing cache to prevent redundant parsing of identical queries
• Performance metrics for GraphQL parsing operations
• Optimized mutation detection and endpoint routing logic
• Improved memory management during GraphQL operations
• Enhanced introspection query handling for better security

HTTP Client Enhancements

Optimized HTTP client configuration with:

• Improved defaults for better performance
• Granular timeout controls via:
  • CLIENT_READ_TIMEOUT
  • CLIENT_WRITE_TIMEOUT
  • CLIENT_MAX_IDLE_CONN_DURATION - Default: 300s
• Connection pool limiting via MAX_CONNS_PER_HOST - Default: 1024
• TLS verification control via CLIENT_DISABLE_TLS_VERIFY
• Improved buffer handling and response processing

Testing & Quality Improvements

Expanded test coverage including:

• Extensive test coverage for all new features
• Integration tests for end-to-end validation
• Dedicated test suites for:
  • Circuit breaker functionality
  • Circuit breaker state transitions
  • Circuit breaker fallback behavior
  • FastHTTP client behavior
  • Memory cache compression
  • Memory cache eviction
  • Gzip error handling

Architectural Impact

These changes significantly improve the proxy's:

• Resilience to backend failures through circuit breaking
• Resource utilization through memory-aware caching
• Performance under load with optimized HTTP and GraphQL handling
• Prevention of cascading failures
• Protection against resource exhaustion