HAProxy One for Kubernetes: Unified traffic management and security

Kubernetes solved deployment. It didn't solve networking. The same platform that lets application teams ship ten times a day leaves infrastructure teams stitching together ingress controllers, cloud load balancers, and per-provider security controls. That patchwork is where production incidents start

You're probably running more than one cluster, more than one ingress controller, and more than one place where security gets enforced. This book is about why that happens, and what the alternative looks like.

This book shows you how to:

• Cut the "mesh tax." Direct-to-pod routing removes the ingress controller and kube-proxy from the request path, so traffic takes one hop instead of three. 

• Name the five problems. Opaque internal routing, external access complexity, configuration drift, multi-cluster silos, and no security at the boundary. 

• Fail over at the pod, not the DNS record. DNS-level failover only knows whether a cluster endpoint answers.

• Stop threats before they reach a pod. Enforce Layer 7 security (including next-gen WAF, bot management, DDoS mitigation, and global rate limiting) at the outer boundary before threats consume cluster resources

• Let developers route without having to be network engineers. Implement controlled developer self-service routing within operator-defined guardrails using familiar annotations and RBAC

• Pick the path that fits, then move at your own pace. Migrate safely from legacy Ingress to Gateway API and external gateways using an incremental, zero-downtime coexistence model