Deploy Custom Apps to a cluster
Deploy a Custom App as a Kubernetes service alongside DIAL Core, then register it so the Unified API routes requests to it. This guide is for DevOps engineers operating a DIAL cluster who have received a Custom App image from an application developer.
A Custom App is a containerized service that DIAL Core calls over HTTP. From an operations perspective, deploying one is the same as deploying any extension: install it with the dial/dial-extension chart and add an entry to DIAL Core's configuration.
Where the full steps live
The end-to-end procedure — building the image, the Docker Compose path for local testing, the dial-extension Helm values, and registering the app in DIAL Core using its in-cluster DNS name — is documented once, on the canonical page:
- Deploy a Custom App — Docker Compose and Helm deployment, with verification steps
This page is not duplicated here; the canonical guide serves both application developers and operators.
Operational checklist
When deploying a developer-supplied Custom App to a shared cluster, confirm:
- Namespace — deploy the app into the same namespace as DIAL Core, or ensure cross-namespace DNS and network policies allow Core to reach it.
- Image registry — the app image is in a registry your cluster can pull from; configure
imagePullSecretsif it is private. - Resource limits — set requests and limits in the
dial-extensionvalues so the app cannot starve DIAL Core. - Registration — add the app to DIAL Core's
applicationsconfiguration using the Kubernetes service DNS name, then reload Core. See Applications configuration. - Access control — restrict the app to the intended roles if it should not be available to all users.
Further reading
- Deploy a Custom App — the full deployment procedure
- Applications configuration — register an app and set its attributes
- Quick Apps installation — enable the Quick Apps feature platform-wide
- Extension chart configuration — workload values, resources, storage, and routing