testing-load-balancers

Validate load balancer behavior including traffic distribution algorithms, health check mechanisms, failover scenarios, session persistence, and SSL termination. Supports testing for NGINX, HAProxy, AWS ALB/NLB, GCP Load Balancers, and Kubernetes Ingress controllers.

• Load balancer deployed and accessible in a test environment • Multiple backend instances running with identifiable responses (hostname headers) • HTTP client tools (curl, wrk, hey, or k6) for sending test traffic • Access to load balancer configuration and health check settings • Ability to stop/start backend instances to simulate failures

Traffic distribution test with curl: `bash #!/bin/bash set -euo pipefail declare -A counts for i in $(seq 1 100); do backend=$(curl -s -H "Host: app.test.com" http://lb.test.com/health \ | jq -r '.hostname') counts[$backend]=$(( ${counts[$backend]:-0} + 1 )) done echo "Traffic distribution:" for backend in "${!counts[@]}"; do echo " $backend: ${counts[$backend]} requests" done ` Failover test sequence: `bash set -euo pipefail

• Verify basic load balancer connectivity: • Send a request through the load balancer and confirm a backend response. • Check the response includes identifying headers (X-Backend-Server, Server) to determine which instance served the request. • Verify SSL/TLS termination works correctly (valid certificate, proper redirect from HTTP to HTTPS). • Test traffic distribution algorithm: • Send 100+ sequential requests and record which backend handled each. • For round-robin: verify even distribution across all backends (within 5% tolerance). • For least-connections: verify the least-loaded backend receives new requests. • For weighted: verify traffic ratio matches configured weights. • Validate health check behavior: • Stop one backend instance. • Verify the load balancer detects the failure within the configured health check interval. • Confirm subsequent requests are routed only to healthy backends (zero errors). • Restart the backend and verify it is returned to the pool after passing health checks. • Test failover scenarios: • Stop all backends except one and verify the remaining backend handles all traffic. • Stop all backends and verify the load balancer returns a 502 or 503 error (not hang). • Simulate slow backend responses and verify timeout behavior. • Validate session persistence (sticky sessions): • Send multiple requests with the same session cookie. • Verify all requests route to the same backend instance. • Verify a new session (no cookie) can route to any backend. • Test connection draining: • Start a long-running request, then remove the backend from the pool. • Verify the in-flight request completes successfully. • Verify new requests route to remaining backends. • Document all results with request/response evidence and timing data.