Kubernetes for AI agents: when code is cheap, proof matters
AI can write code faster than teams can verify it. Build a Kubernetes platform for parallel QA, disposable clusters and evidence you can trust.
A field guide for platform engineers
Kubernetes runs applications across a group of servers. We explain the tools around it: what they do, why you might need them and how to choose between the alternatives.
Fresh from Kubedex
New articles and substantial updates, with a quick look at what changed.
AI can write code faster than teams can verify it. Build a Kubernetes platform for parallel QA, disposable clusters and evidence you can trust.
Need to store uploads, backups or application files on your own servers? See how MinIO, Garage, SeaweedFS and Ceph compare.
Argo CD deploys applications from Git. Learn how to manage many deployments and recover when a release goes wrong.
Collect application logs, metrics and traces in one place. Understand the Collector, Grafana Alloy and how to avoid duplicate data.
Run language models on Kubernetes. See what vLLM, KServe and llm-d each do and when you need them.
Help machine learning jobs share scarce GPUs. Understand allocation, job queues and the choices for sharing a device.
Find known security flaws in container images before deployment. Compare Trivy and Grype, then build a useful scanning process.
Check who published a container image before running it. Learn how signing and cluster admission policies work together.
Replacing Bitnami packages? Understand which application images, installation charts and stored data need to move.
Move a PostgreSQL database from Bitnami packaging to CloudNativePG, with a tested restore example and checks for your data.
Add and remove AWS servers as application demand changes. Compare Karpenter, Cluster Autoscaler and EKS Auto Mode.
Replace the retired ingress-nginx controller that routes web traffic to applications. Compare replacements and plan the move.
Start here
Compare Argo CD and Flux for deploying applications from Git: dashboards, Helm releases, access controls and recovery after a bad change.
Compare Amazon EKS, Azure AKS and Google GKE: what each managed Kubernetes service provides and which fits your applications and cloud services.
Compare Istio and Linkerd for encrypting traffic between services, controlling access and investigating requests across applications.
Navigate the ecosystem
From the first hop to the last backup,
follow the problem you’re solving.
Get web requests to the right service and connect applications safely.
Deploy application changes and keep environments up to date.
See whether applications are healthy and investigate problems.
Check software, protect credentials and control access.
Choose where applications run and add capacity when they need it.
Store application data, keep backups and recover from failures.
From the publication
Good recommendations show their work.
Sources, review dates and honest limits. That’s the Kubedex standard.