Skip to main content

Running Ignition on Kubernetes

Before continuing

Kubernetes runs the same inductiveautomation/ignition container you already know, but it adds the machinery to schedule that container across a cluster, give it durable storage, reach it over a stable address, and feed it configuration and secrets. This pillar covers the pieces of that machinery that matter for Ignition specifically, and the operational patterns that come up once a gateway is running on a cluster.

The official Helm chart at charts.ia.io assembles most of these resources for you, so these guides focus on the why behind the chart's choices and on the integration work the chart leaves to you.

What's here​

  • Kubernetes concepts for Ignition: the cluster primitives the Helm chart relies on (StatefulSets, PVCs, Services, ConfigMaps, Secrets) and why Ignition needs each one.
  • External modules from S3: mount third-party .modl files from an S3 bucket into all gateways in the namespace without baking them into pods.
  • External Secrets: wire AWS Secrets Manager (or any provider) to Kubernetes Secrets via the External Secrets Operator, covering the license key, git credentials, and API key.
  • GitOps with ApplicationSets: use ArgoCD's matrix generator and a config.yaml discovery marker to deploy charts across clusters automatically.
  • Config Sync: keep Ignition project files and gateway config in sync with Git using the Stoker operator's GatewaySync resource.

Short, one-screen operational tips will be collected here under a Tasks sub-section as the pillar grows, following the Kubernetes documentation Concepts / Tasks / Tutorials / Reference convention.