[FEATURE]: Centralized Redis configuration

[crivetimihai]

Summary

Add Redis connection pool and leader election settings to config.py and .env.example to eliminate hardcoded values scattered across the codebase.

Problem: Hardcoded Values

Setting	Hardcoded Location	Current Value
decode_responses	llmchat_router.py:65, oauth_manager.py:66, event_service.py:203	True
socket_connect_timeout	oauth_manager.py:66	5
socket_timeout	oauth_manager.py:66	5
leader_ttl	gateway_service.py:347	40
max_connections	(not configured)	—
health_check_interval	(not configured)	—

Solution: Add to config.py

Add after existing Redis settings (around line 1098):

# Redis Configuration - Performance Optimized

# Connection Pool Settings
redis_decode_responses: bool = Field(default=True, description="Return strings instead of bytes")
redis_max_connections: int = Field(default=50, description="Connection pool size per worker")
redis_socket_timeout: float = Field(default=2.0, description="Socket read/write timeout in seconds")
redis_socket_connect_timeout: float = Field(default=2.0, description="Connection timeout in seconds")
redis_retry_on_timeout: bool = Field(default=True, description="Retry commands on timeout")
redis_health_check_interval: int = Field(default=30, description="Seconds between connection health checks (0=disabled)")

# Leader Election Settings
redis_leader_ttl: int = Field(default=15, description="Leader election TTL in seconds")
redis_leader_key: str = Field(default="gateway_service_leader", description="Leader key name")
redis_leader_heartbeat_interval: int = Field(default=5, description="Seconds between leader heartbeats")

Solution: Add to .env.example

# Redis Connection Pool - Performance Tuned
REDIS_MAX_CONNECTIONS=50
REDIS_SOCKET_TIMEOUT=2.0
REDIS_SOCKET_CONNECT_TIMEOUT=2.0
REDIS_RETRY_ON_TIMEOUT=true
REDIS_HEALTH_CHECK_INTERVAL=30
REDIS_DECODE_RESPONSES=true

# Redis Leader Election - Multi-Node Deployments
REDIS_LEADER_TTL=15
REDIS_LEADER_HEARTBEAT_INTERVAL=5
REDIS_LEADER_KEY=gateway_service_leader

Performance Optimization Rationale

Setting	Default	Rationale
max_connections	50	Async apps need more connections; 50 handles ~500 concurrent requests per worker
socket_timeout	2.0s	Fast failure detection; Redis ops should complete in <100ms
socket_connect_timeout	2.0s	Quick connection failure; don't block event loop waiting
health_check_interval	30s	Periodic pool health check prevents stale connections
leader_ttl	15s	Faster failover (was 40s); 15s balances speed vs false positives
leader_heartbeat_interval	5s	Refresh 3x before TTL expires (15/5=3); ensures leader doesn't lose lock

Acceptance Criteria

• All 9 new settings added to config.py with Field definitions
• All settings documented in .env.example with comments
• Settings use performance-optimized defaults
• make check-env passes