Skip to main content
Version: Current

Flink MiniCluster

This variant runs streaming scenarios on a Flink MiniCluster inside the Designer JVM. Request-Response scenarios run on the embedded Lite engine in that same pod. No separate Flink cluster is required. It is the quickest evaluation setup; both embedded engines restart with Designer.

Values​

Save this example as values.yaml. It uses local port-forwarding and the bundled database, Kafka and monitoring services. Follow the common requirements and installation steps.

# Local evaluation: access Designer and scenario endpoints with kubectl port-forward.
ingress:
enabled: false
global:
publicUrls:
requestResponse: http://localhost:8191
gateway: http://localhost:18147
imagePullSecrets:
- name: nussknacker-registry
license:
existingSecret: nussknacker-license
environment: trial
nussknacker:
streaming:
deploymentTargets:
- name: Flink MiniCluster
useMiniClusterForDeployment: true
requestResponse:
deploymentTargets:
- name: Request-Response embedded

The MiniCluster's Flink version comes from the Designer image. Its slots, managed memory and local checkpoint paths can be changed under nussknacker.streaming.deploymentConfig.miniCluster.config. Increase Designer resources before trying larger workloads. The default storage paths use the Designer PVC; they do not configure storage for any remote Flink cluster.

Install and verify​

Create the registry/license Secrets and install the chart. Then open Designer, create or import a streaming scenario, and deploy it to Flink MiniCluster. Check that the scenario reaches RUNNING and processes a record. Request-Response scenarios deploy to Request-Response embedded.

For production streaming workloads, use an appropriate external Flink cluster or Ververica Platform.