# Kubernetes, without the cluster babysitting. Kubernetes on EKS Create a production-ready EKS cluster in your own AWS account, or bring the one you already run. Deploy web services, workers, and any Helm chart from a form or a typed config. Ravion wires up Karpenter, autoscaling, secrets, load balancers, logs, and metrics, and tells you exactly why a pod won't start. [Start free](https://app.ravion.com/signup) [Book a demo](https://www.ravion.com/demo) Running production on AWS for engineering teams since 2021. ## Getting a cluster is easy. Running one is the job. Most teams don't choose Kubernetes for the YAML. They choose it for the ecosystem, then spend a platform engineer's year keeping it alive. ### Day 2 never ends Version upgrades, node pools, autoscalers, ingress controllers, secrets, and an observability stack, each with its own Helm chart and upgrade path. ### Every deploy is a wall of YAML Developers copy manifests between services and wait on the one person who knows how the cluster fits together. ### When a rollout breaks, you're in kubectl CrashLoopBackOff, a failing probe, a pending pod. Finding out why means describe, logs, and events across three terminals. With Ravion ## The power of Kubernetes. The workflow of a PaaS. From an empty AWS account to debugging a live rollout, in one place. Step 1 ### Create an EKS cluster, or bring yours Start a production-ready cluster with private networking, encrypted secrets, control plane logging, Pod Identity, and the core EKS add-ons, from an open source Terraform module you can read. Already run EKS? Connect it with its name, endpoint, and an IAM role, and deploy to it the same way. [See the cluster module](https://www.ravion.com/docs/module-definitions/catalog/rvn-eks-cluster) Step 2 ### Switch on the add-ons you'd otherwise maintain On a new cluster or one you bring: Karpenter, External Secrets, and EBS storage are on by default. Turn on KEDA, cert-manager, shared load balancers, and logs, metrics, and traces that stay in your account or go to Datadog, Grafana Cloud, New Relic, and more. [See the add-ons module](https://www.ravion.com/docs/module-definitions/catalog/rvn-eks-addons) Step 3 ### Deploy services without writing manifests Describe a web service or worker once: image or Dockerfile, CPU and memory, replicas, autoscaling, domain, secrets, and IAM role. Ravion renders the Deployment, Service, HPA, PodDisruptionBudget, and load balancer binding for you. [See the web service module](https://www.ravion.com/docs/module-definitions/catalog/rvn-eks-web) Step 4 ### Push to deploy, with Helm releases you can trust Every deploy is a Helm release with a values diff, live progress, and an automatic rollback if pods never become ready. Cancel a rollout or roll back to any earlier revision in one click. Step 5 ### Know why a pod is failing, without kubectl An in-cluster agent reports every pod, Kubernetes event, and crash log back to Ravion. When a rollout stalls, you see the diagnosis, the last lines the container printed, and the suggested fix, right next to the deploy that caused it. ## Every kind of workload, one module away. Each module is a form for developers and a typed config for agents, backed by open source Terraform and Helm. ## Your cluster. Your account. Our autopilot. The cluster, nodes, load balancers, and observability data are plain AWS resources in your account. Leave any time and keep running them with Terraform and Helm. Interact with Ravion through the UI, CLI, API, or Git. Ravion sends releases to your AWS account and gets pods and events back. Everything else runs inside your account. - UI - CLI - API - Git RavionControl plane - Push-to-deploy Helm releases and rollbacks - Workload, pod, and event view per deploy - Approvals and audit history for people and agents Releases Pods and events Your AWS accountYou own it Your VPCPrivate network EKS clusterEKS - web - api - worker - chart Karpenter nodes, On-Demand, Spot, or Fargate Wired in for you - Shared ALBs and NLBs - Secrets Manager via External Secrets - Loki, Prometheus, and Tempo on S3 - Postgres and MySQL - Redis - Your IAM, bill, and credits ## Everything a production cluster needs ### New cluster or bring your own. Create a cluster from the open source module, or connect an EKS cluster you already run and deploy to it with the same modules, add-ons, and agent. ### Karpenter and autoscaling. Right-sized nodes from Karpenter with Spot interruption handling, plus HPA on CPU and memory and KEDA triggers for bursts. ### Spot, On-Demand, or Fargate. Choose per workload. Keep a baseline on On-Demand and burst onto Spot when traffic spikes. ### Any Helm chart. Install charts from Helm repositories, OCI registries like ECR and GHCR, or Git, with full control of values and Pod Identity. ### Atomic releases and rollbacks. Failed rollouts roll back on their own. Roll back to any earlier Helm revision, or restart a workload, from the dashboard or CLI. ### Logs, metrics, and traces in your account. Loki, Prometheus with Thanos, and Tempo backed by S3, or ship to CloudWatch, Datadog, Grafana Cloud, New Relic, Splunk, or any OTLP endpoint. ### Domains and load balancers. Attach services to shared public or private ALBs with WAF, custom domains, and Ravion-managed domains. ### Drift and outside deploys. See when someone changes a workload with kubectl or Helm outside Ravion, and exactly what changed. ### Agent-ready, behind approvals. Agents deploy, roll back, and read logs through the CLI and API, with the same approvals and audit history as your team. ## Bring your workloads over one service at a time. Run Ravion's cluster beside the one you have, then shift traffic service by service. Porter runs your apps on a cluster it manages behind its own abstractions. Teams move to Ravion to keep a managed experience while owning every module, chart, and value that runs in their AWS account. PorterRavion Porter clusterrvn-eks-cluster and rvn-eks-addons Web servicervn-eks-web Worker servicervn-eks-worker Helm add-onrvn-eks-chart porter.yamlProject config (ravion.yaml) Move in three steps 1. 01 Create an EKS cluster with Ravion in the same AWS account. 2. 02 Add a module for each Porter service and deploy it while Porter keeps serving traffic. 3. 03 Shift traffic with DNS, then remove the services from Porter. Kubernetes questions ### Do I need to know Kubernetes? No. Developers deploy web services and workers from a form or ravion.yaml, and Ravion renders the Kubernetes objects and Helm release. When you want to go deeper, every chart, value, and module is open source and yours to change, and kubectl works as usual. ### Should I use EKS or ECS? Both run in your AWS account with the same push-to-deploy workflow. Choose ECS if you just want containers to run with the least to operate. Choose EKS when you need the Kubernetes ecosystem: Helm charts, operators, Karpenter, or a team that already knows Kubernetes. You can run both in the same project. ### Can I deploy my own Helm charts? Yes. The EKS Helm Chart module installs any chart from a Helm repository, an OCI registry such as ECR, GHCR, or Docker Hub, or a Git repository, with full control of values, registry credentials, and an IAM role through Pod Identity. ### Can we use the EKS cluster we already run? Yes. Point web services, workers, Helm charts, and add-ons at an existing EKS cluster by entering its name, ARN, API endpoint, and certificate, plus an IAM role with a cluster access entry that Ravion's deploy runner assumes. Turn on Ravion EKS Management to install the agent for pod, event, and log visibility. Upgrades and node groups of a cluster you bring stay with whoever manages it today, or create a new cluster with Ravion to hand those over too. ### What runs inside my cluster? An optional agent reports workloads, pods, Kubernetes events, and crash logs back to Ravion over an outbound connection, so nothing needs to reach into your cluster. It is read-only by default. ### What happens if we leave Ravion? You keep everything. The cluster, nodes, load balancers, and data are plain AWS resources built from open source Terraform modules, and your workloads are ordinary Helm releases. Keep running them with Terraform and Helm. ## Keep Kubernetes. Lose the busywork. Create your first production-ready EKS cluster in your AWS account today. [Start free](https://app.ravion.com/signup) [Book a demo](https://www.ravion.com/demo) Free to start / No credit card / Runs in your AWS account