kubectl-config-doctor is a CLI plugin that detects broken, misconfigured, or unreachable kubeconfig setups—catching issues like duplicate contexts, invalid auth methods, missing files, or unreachable APIs.

More: https://ku.bz/Whfwxjb38

KubeDesk is a cross-platform Kubernetes desktop app for browsing resources, viewing health and topology, streaming logs, opening shells, and making dry-run validated changes.

More: https://ku.bz/-x0txgkcy

EKS Hybrid Nodes addresses almost every on-prem Kubernetes pain point — without requiring a full cloud migration.

Raglin Anthony explains how: AWS manages the control plane (patching, scaling, hardening), while your worker nodes stay on-premises. EKS Pod Identity and IAM roles for service accounts handle identity natively. Networking reuses your existing VPN or Direct Connect setup. And nodeadm handles node upgrades.

You keep control of your infrastructure. AWS takes the control plane burden.



Watch the full interview: https://ku.bz/2XqMJnLVx

This article explains three things that make DNS feel strange inside Kubernetes on AWS: the ndots setting, NodeLocal DNS caching, and a hidden per-ENI packet limit.

More: https://ku.bz/SPFB3xHDf

Ben Poland, Senior Staff Platform Engineer @ Faire, discusses the unexpected Kubernetes scaling challenges his team encountered when migrating their CI system from Jenkins to Buildkite.

He explains how their new architecture created thousands of smaller pods instead of fewer large ones, which exposed internal Kubernetes throttling mechanisms they hadn't anticipated.

The team also hit etcd space limitations due to the sheer number of concurrent pods running in their managed Kubernetes clusters.

Watch the full episode: https://ku.bz/klBmzMY5-

SnowOps Labs is a Kubernetes practice lab that builds a local cluster, injects realistic failures, and grades your fixes through guided scenarios and incidents.

More: https://ku.bz/k8HVXFY_V

Owumi Festus, Founding AI Engineer @ Plexe AI at Plexe AI, explains the specific architectural decisions he made when building a lightweight version of Kubernetes from scratch.

Key simplifications include:

- Replacing etcd with in-memory storage to avoid distributed consensus mechanisms and high availability requirements
- Removing containers entirely and using state logging instead to focus on pod lifecycle understanding without container runtime complexity
- Eliminating authentication, service discovery, and RBAC to reduce permission overhead
- Using polling instead of watches and streams for API server communication
- Limiting resource definitions to pods only rather than supporting deployments and other complex resources

Watch the full episode: https://ku.bz/pf5kK9lQF

This article explains why a ConfigMap you just changed via the API still appears outdated in the mounted file, and why the kubelet can take up to a minute to catch up.

More: https://ku.bz/NtynkHvDl

CPU and memory usage need context before you change Kubernetes requests and limits.

Throttling, pressure, and runtime events help explain the usage graph. Looking at each replica separately shows differences that a workload average can hide.

We are releasing the Kubernetes Resource Review Mixin, an open-source Grafana dashboard with Prometheus recording rules.

You can use it to:

- See how each Pod's CPU and memory usage compares to its requests and limits.
- Inspect CPU throttling, CPU and memory pressure when PSI metrics are available, and OOM and restart activity.
- Review replicas and workload revisions separately, with p95, maxima, and resource differences in the tables.

The mixin is an extension of our book on rightsizing Kubernetes workloads, The Technical Guide to Kubernetes Rightsizing.

Clone it here: https://github.com/learnk8s/resource-review-mixin

Hortator lets AI agents spawn sub-agents at runtime, with each agent running in its own pod with budget caps, network policies, PII redaction, and capability inheritance so children can never escalate beyond their parent's permissions.

More: https://ku.bz/KqYBl17Gv

LearnKube Day includes ten five-minute lightning talks from Kubernetes practitioners sharing practical lessons with the community.

Graziano Casto, Developer Advocate at Akuity, is one of the featured lightning-talk speakers at LearnKube Day Salt Lake City 2026.

The event brings together, lightning talks, YAML Games, and a community reception with a live DJ(our own Amin Astaneh).

Learn more: https://learnkube.com/learnkube-day-salt-lake-city-2026

This week on Learn Kubernetes Weekly 206:

🌐 DNS Is Weird Inside Kubernetes on AWS
⏳ TIL: Kubernetes ConfigMap Volume Mounts Are Eventually Consistent — Not Immediately Consistent
🚨 CVE-2025-69902: OS Command Injection in kubectl-mcp-tool
🐘 Building the Foundation for PostgreSQL Lifecycle Management in a DBaaS on Harvester
💰 We Cut Our Cassandra Bill From $30K to $10K a Month by Self-Hosting on Kubernetes

Read it now: https://kube.today/issues/206

⭐️ This issue is brought to you by Teleport — stop sharing credentials and give every engineer, service, and AI agent a login that expires automatically https://ku.bz/2tJxlTyDy

kubectl-netshoot is a kubectl plugin that spins up a netshoot container in a pod or on a node, so you can debug cluster networking without adding any tools to your own images.

More: https://ku.bz/2N04Jyj70

AI review improves when it stops acting like a line-by-line linter.

Shivay Lamba of Qualcomm explains that modern AI review tools work best when they inspect the repository as a connected system, relate pull request changes to surrounding code, and use engineering guidance that reflects how the team actually writes software.





Watch the full interview: https://ku.bz/lzl3pF_Ft

The AWS team built a three-step ramp for running AI on Kubernetes: infrastructure, blueprints, and optimization.

Tsahi Duek walks through the AI on EKS open source project: start with an inference-ready cluster template, deploy a model with vLLM using a pre-built blueprint, then optimize — shrinking cold starts, benchmarking GPU types, and using MIG or DRA to pack more workloads per GPU.





Watch the full interview: https://ku.bz/2r41YKBZb

This tutorial shows how to run a fully local RAG system on Kubernetes with PostgreSQL and pgvector, local embeddings, and a CPU-based Qwen model without API keys.

More: https://ku.bz/0-0X6mLS9

This article argues that uptime, latency and error rate only tell you a system is alive, not correct, and shows which behaviour-level signals and baselines belong in dashboards and alerts instead.

More: https://ku.bz/v09SHsxGG

Kubernetes abstracts infrastructure, and that's exactly why finding the biggest pod talkers has been so hard.

Elamaran Shanmugam explains the core problem: infrastructure speaks IPs and ENIs, Kubernetes speaks pods and namespaces, and there's no mechanism to bridge the two. With container network observability, a network flow monitor agent runs on every node, providing enriched pod-to-pod communication metrics for your EKS environment and Grafana dashboards.



Watch the full interview: https://ku.bz/DCxhgWQqS

Jorrick Stempher explains how his team designed realistic load testing for their Kubernetes predictive scaling system.

He details their approach using virtual users that send requests to multiple API endpoints and reveals a critical insight about avoiding instant load spikes that cause immediate failures.

The discussion covers how they implemented ramp-up periods of around 5 seconds to simulate realistic traffic patterns, since jumping from zero to thousands of requests per second instantly crashes nodes and doesn't reflect real-world usage.

Watch the full episode: https://ku.bz/clbDWqPYp

DockLog is a lightweight, self-hosted Docker log viewer for teams, with real-time streaming, RBAC and audit logging, plus optional Kubernetes visibility.

More: https://ku.bz/5Lr7Mj09-