First ELB

Load balancer.

Overview

Your first ELB (Elastic Load Balancer) is the moment traffic distribution moves from theory to production. AWS offers three flavours; the team picks one, wires it correctly, and inherits a load-balancing fabric that scales for years.

• Three flavours. ALB for HTTP/HTTPS at layer 7, NLB for TCP/UDP at layer 4 with ultra-low latency, GLB for third-party appliance traffic. The pick is per workload, not per habit.
• Health checks. The ELB removes unhealthy targets automatically. Deliberate configuration matters; bad checks produce flapping or false-healthy targets that take outages.
• TLS termination plus target groups. The ELB terminates TLS using ACM certificates and routes to target groups (EC2, IP, Lambda, ECS, EKS). Cert management centralises; routing organises around target groups rather than individual hosts.
• Auto scaling integration. ELBs integrate with auto scaling groups. New instances register automatically as they launch; capacity becomes self-managing rather than a checklist.

The approach

The practical approach is start-with-ALB-for-HTTP, declare everything as code, and tune health checks deliberately. The first ELB is the template; subsequent ones follow it without rediscovery.

• ALB for HTTP/HTTPS. Most workloads are HTTP. ALB is the default; path-based routing, host-based routing, and redirects come for free at layer 7.
• NLB for low-latency or non-HTTP. NLB handles TCP/UDP, ultra-low latency, and static IPs. Use it only when ALB does not fit; the operational shape is heavier.
• Terraform or CDK plus ACM. ELB and target groups are infrastructure as code; click-ops produces drift. ACM certs are free, auto-renew, and integrate natively for every public-facing ELB.
• Tune health checks. Path, threshold, interval, and timeout each matter. Defaults rarely fit; bad checks produce flapping during deploys or false-healthy targets during real outages.

Why this compounds

The first ELB compounds. Each subsequent ELB inherits the patterns, the Terraform modules, and the operational runbooks; team velocity grows because nobody re-derives the same setup.

• Reusable Terraform modules. The module for ALB plus target group plus listener becomes standard. New services slot into the pattern in hours rather than days.
• Established TLS management. ACM certs and ELB integration become the team's default. New public-facing services inherit cert hygiene without custom work.
• Shared observability. ELB CloudWatch metrics (5xx rate, latency, target health) become standard dashboards. Investigation across services follows a consistent shape.
• Operational maturity plus compounding habit. Each ELB produces lessons in health-check tuning, capacity planning, and security-group discipline. Year one lays the foundation; year two onwards the patterns are reflexive.

The first ELB is one of those infrastructure investments that pays off across years of operation. Nova AI Ops integrates with ELB metrics, surfaces patterns across services, and supports the team's traffic distribution discipline.