Kubernetes Networking Deep Dive

Overview

Explain the Pod network (CNI), CoreDNS service discovery, kube-proxy modes, and how NetworkPolicies enforce east-west rules.

Every Pod gets an IP via CNI (Calico, Cilium, kindnet…). CoreDNS answers svc.ns.svc.cluster.local. NetworkPolicies are enforced by the CNI plugin — not by kube-apiserver alone.

This is a core tutorial in Module 11 · Networking Deep Dive of the REBASH Academy Kubernetes for Cloud & DevOps Engineers series — written for Cloud, DevOps, Platform, and SRE engineers.

Prerequisites

• Services · Security Hardening

Learning Objectives

By the end of this tutorial, you will be able to:

• Name your cluster’s CNI
• Resolve Service DNS from a debug Pod
• Outline iptables vs IPVS vs eBPF
• Write a simple NetworkPolicy

Architecture

This topic’s control points and relationships are shown below.

Theory

What it is

Kubernetes networking rests on a few guarantees: every Pod gets an IP; Pods can reach each other without NAT (within the cluster model); agents on nodes implement that fabric via a CNI plugin. CoreDNS provides service discovery. kube-proxy (iptables/IPVS) or eBPF dataplanes implement Services. NetworkPolicies express allow/deny rules enforced by the CNI — not by the API server alone.

Why it matters

Most “it works on my laptop” failures in production are DNS, NetworkPolicy, or CNI MTU/routing issues. DevOps engineers who can name their CNI, query CoreDNS, and read EndpointSlices debug faster than those who only restart Pods. Security teams need NetworkPolicies that actually enforce.

How it works (mental model)

1. CNI assigns Pod IPs and programmes routes/overlays/eBPF maps.
2. kubelet and runtime attach the Pod to the network namespace.
3. Services select Pods; EndpointSlices list ready backends; the dataplane DNAT/load-balances to Pod IPs.
4. Pods resolve service.namespace.svc.cluster.local via CoreDNS (kube-dns Service).
5. NetworkPolicy objects are watched by the CNI agent; non-matching traffic is dropped when policies select a Pod.

Flat Pod network + Services + DNS is the mental model; overlays and cloud routing are implementation details.

Key concepts / comparisons

Layer	Component
Pod IP fabric	CNI (Calico, Cilium, kindnet, …)
Service VIP	kube-proxy / eBPF
DNS	CoreDNS
Policy	NetworkPolicy (+ CiliumNetworkPolicy etc.)

kube-proxy mode	Trait
iptables	Common default
IPVS	Better scale characteristics
eBPF (Cilium)	Often replaces kube-proxy

Common pitfalls

• Assuming NetworkPolicies work without a supporting CNI — they become no-ops.
• DNS failures from CoreDNS Pending/CrashLoop — check kube-system first.
• Debugging Service traffic without checking endpoints emptiness.
• Overlapping NetworkPolicies that unintentionally isolate CoreDNS (egress to DNS must remain).
• Confusing NodePort exposure with Pod network reachability from outside.

Hands-on Lab

Create a workspace for this tutorial.

mkdir -p ~/rebash-k8s/module-11 && cd ~/rebash-k8s/module-11

Focus: Trace ClusterIP Service DNS and Pod-to-Pod connectivity

Step 1 – Create two Pods and a Service

kubectl create namespace rebash-lab
kubectl -n rebash-lab run server --image=nginx:1.27-alpine --port=80 --labels=app=server
kubectl -n rebash-lab expose pod server --port=80 --name=server
kubectl -n rebash-lab run client --image=busybox:1.36 --restart=Never --command -- sleep 3600
kubectl -n rebash-lab wait --for=condition=Ready pod/server pod/client --timeout=90s

Step 2 – Resolve DNS and curl across the Service

kubectl -n rebash-lab get svc server -o wide
kubectl -n rebash-lab exec client -- nslookup server.rebash-lab.svc.cluster.local
kubectl -n rebash-lab exec client -- wget -qO- http://server.rebash-lab.svc.cluster.local/ | head -n 5
kubectl -n rebash-lab get endpoints server -o yaml | head -n 30

Final step – Cleanup note

kubectl delete namespace rebash-lab --ignore-not-found
# Workspace kept for notes; remove with: rm -rf "$(pwd)" when finished

Validation

• Lab commands run under ~/rebash-k8s/module-11/
• You can explain each Theory section in your own words
• You used modern tooling where it applies to this topic
• You can describe one production failure mode for this topic

Code Walkthrough

Production practice for Kubernetes Networking Deep Dive always combines:

1. Inspect before you change (status, plan, logs, dry-run)
2. Prefer reversible, documented changes (Git, IaC, drop-ins, version pins)
3. Capture evidence (command output, pipeline logs) for handovers
4. Prefer current tools and APIs over legacy shortcuts
5. Least privilege — escalate credentials only when required

Keep runbooks short enough to follow under pressure. Automate checks; keep humans for judgement.

Security Considerations

• Treat credentials and tokens for kubernetes as privileged — never commit them
• Prefer short-lived auth (OIDC, roles, SSO) over long-lived keys
• Validate blast radius before apply/deploy/delete operations
• Restrict who can approve production changes
• Collect audit logs; limit who can read sensitive traces

Common Mistakes

Assuming NetworkPolicies work without a supporting CNI — they become no-ops.

Validate assumptions against the Theory section and official docs before changing production.

DNS failures from CoreDNS Pending/CrashLoop — check kube-system first.

Lab shortcuts (open security groups, admin roles, skip approvals) must not ship unchanged.

Changing production without a rollback path

Always know how to revert (previous artefact, prior release, state rollback, DNS failback).

Best Practices

• Encode Kubernetes Networking Deep Dive changes as code and review them in pull requests
• Pin versions (images, modules, actions, provider plugins)
• Separate environments with clear promotion gates
• Alert on symptoms with runbooks attached
• Destroy lab resources; tag everything with owner and expiry where possible

Troubleshooting

Symptom	Likely cause	Fix
Auth / permission denied	Wrong identity, policy, or scope	Check caller identity, roles, and least-privilege policies
Timeout / no route	Network, DNS, security group, or endpoint	Trace path, DNS, and allow-lists before retrying
Drift / unexpected plan	Manual change or wrong state/workspace	Reconcile desired vs actual; avoid click-ops on managed resources
Pipeline/job red	Flaky step, cache, or missing secret	Read failing step logs; bisect recent workflow/config changes
Cost spike	Idle load balancer, NAT, oversized compute	Inventory billable resources; stop/delete labs promptly

Summary

Kubernetes Networking Deep Dive is essential for Cloud and DevOps engineers working with kubernetes. Practise the lab until the inspection and change path is muscle memory, then continue the track.

Interview Questions

1. How does Pod networking typically work regarding IP addresses?
2. How does CoreDNS resolve a Service name inside a cluster?
3. What is the difference between ClusterIP, NodePort, and LoadBalancer?
4. How can NetworkPolicy restrict east-west traffic, and what must the CNI support?
5. What symptoms suggest a CNI or kube-proxy problem rather than an application bug?

Sample answer — question 2

Services get a stable DNS name like name.namespace.svc.cluster.local that resolves to the ClusterIP. kube-dns/CoreDNS answers these queries for in-cluster clients.

Sample answer — question 4

NetworkPolicy only enforces if the CNI implements it. Policies default-deny unused paths, allow needed namespaces/pods/ports, and should be tested so you do not lock out DNS or probes accidentally.

Related Tutorials

• Course overview
• Monitoring and Logging in Kubernetes

References

• Cluster networking · DNS for Services