Load balancing: traffic and resilience

Objectives and progression

A technical course with six modules and fictional APS cases. Interpret L4/L7, TLS boundaries, algorithms, affinity, checks, retries, per-worker limits, post-failure capacity, and target retirement. Distinguish NGINX, HAProxy 3.2, Kubernetes, and AWS ALB behavior using primary references. Internal assessment of 24 decisions in 60 minutes without external certification or executed labs.

Audience: APS L2/L3 teams, middleware administrators, platform teams, and technical managers.

Prerequisites: TCP/IP, HTTP/HTTPS, and TLS fundamentals; cases provide required limits and observations.

300 estimated study minutes

• Identify what the load balancer sees and which dependencies remain on the path.
• Choose distribution signals and distinguish persistence from data recovery.
• Define what each check establishes and what failure causes.
• Control repetition, duration, concurrency, and trust in received metadata.
• Calculate headroom after failures and recognize load amplification.
• Validate actual traffic and coordinate request completion during changes.

Modules

1. Routing and TLS boundaries
2. Algorithms, affinity, and state
3. Health checks and readiness
4. Retries, limits, and client origin
5. Capacity and cascading failures
6. Draining, releases, and operations

Continue learning

Technical assessments

• HTTP / HTTPS
• TLS and Certificates
• Middleware
• Application Production Support

References and version

DR load balancing 2026-09; selected NGINX, HAProxy 3.2, Kubernetes and AWS ALB behavior

• NGINX HTTP load-balancing methods · 2026-09-30
• NGINX HTTP upstream behavior and version notes · 2026-09-30
• NGINX transport-stream upstreams · 2026-09-30
• NGINX trusted client-address replacement · 2026-09-30
• HAProxy 3.2 configuration manual · 2026-09-30
• Kubernetes startup, liveness and readiness probes · 2026-09-30
• Kubernetes pod termination and endpoint conditions · 2026-09-30
• NGINX TLS ClientHello preread · 2026-09-30
• NGINX active upstream health checks · 2026-09-30
• AWS ALB target connections · 2026-09-30
• AWS ALB health and fail-open behavior · 2026-09-30
• Google SRE datacenter load balancing · 2026-09-30
• Google SRE handling overload · 2026-09-30
• AWS ALB algorithms and deregistration · 2026-09-30
• NGINX retries, timeouts and upstream forwarding · 2026-09-30
• NGINX virtual-server and request selection · 2026-09-30
• HTTP semantics and retry conditions · 2026-09-30