The Best Kubernetes Feature You Forgot About

Hey there,

If you’re losing track of numbers with NodePort, and LoadBalancer is burning through your cloud budget, you’re going to like this one. 

In today's issue: a quick hit on Ingress and why it might be worth a second look, plus a few links and a timeless thread worth your scroll.

Let’s dive in.

Most teams jump straight to those options for external access, but neither scales gracefully. With NodePort, you’re juggling ports across nodes (30080 here, 30443 there), and Load Balancer often means spinning up (and paying for) a separate cloud load balancer per service.

They also fall short on the routing intelligence modern applications need. No path-based routing, no built-in SSL, no advanced strategies like canary deployments or A/B testing.

That’s where Kubernetes Ingress shines. A single entry point for external traffic, with built-in TLS and Layer 7 routing that can cut costs and simplify your architecture. It also unlocks smarter traffic patterns, like routing /api to one service and /shop to another, without juggling dozens of exposed ports.

Ingress itself is just the rules. The real work is handled by the Ingress controller (NGINX, Traefik, HAProxy, etc.), which enforces those rules and directs the traffic.

If you’re running multiple services, planning experiments like canary rollouts, or just tired of the NodePort dance.

Popular controllers to consider:

• NGINX – battle-tested and widely adopted

• Traefik – developer-friendly and flexible

• HAProxy – a performance powerhouse

 To get grounded in Kubernetes Ingress, check out the full guide here 

The Register: “If you can't beat the AI, let it live inside you”. Corporate wellness meets AI strategy: one of you’s getting replaced, and it’s not the chatbot. Read it here. 

Reddit (Timeless): "Why did YOU transition from software dev to devops?". Top comment: "My therapist said I needed to cry more.” Still hits. Read the thread.

InfoQ: “Pinterest retires Hadoop for Moka, a Kubernetes-native data platform.” Spark on EKS, ARM savings, and YuniKorn scheduling. Hadoop who? Read it here.

The New Stack: “What K8s Users Really Think About AI in 2025.” Not hype, but real workloads, real bills, and real headaches. 90% expect AI/ML on Kubernetes to grow this year, with operators scrambling to keep up. Read it here.

And it’s a wrap!

See you Friday for the week’s news, upcoming events, and opportunities.

If you found this helpful, share this link with a colleague or fellow DevOps engineer.

Divine Odazie
Founder of EverythingDevOps