nodestream-plugin-k8s 0.15.0

A plugin for nodestream that manages k8s resources.

Nodestream Kubernetes Plugin

NOTE: This plugin is still in development and is not ready for production use.

Overview

This plugin allows you to deploy nodestream on a Kubernete cluster. As your project evolves by adding more data pipelines, nodstream-plugin-k8s will be able to manage your kubernetes resources for you.

Installation

pip install nodestream-plugin-k8s

Usage

1. Create a nodestream.yaml file and project.

# nodestream.yaml
scopes:
   crons: 
      pipelines:
        - path: pipelines/crons/my_scheduled_pipeline.yaml
  perpetual:
      pipelines:
        - path: pipelines/perpetual/my_kafka_pipeline.yaml

2. Annotate your pipeline with nodestream-plugin-k8s annotations.

# nodestream.yaml
scopes:
   crons:
      pipelines:
        - path: pipelines/crons/my_scheduled_pipeline.yaml
          annotations:
            nodestream_plugin_k8s_schedule: "0 0 * * *"
    perpetual:
       pipelines:
          - path: pipelines/perpetual/my_kafka_pipeline.yaml
            annotations:
              nodestream_plugin_k8s_conccurency: 1

3. Run nodestream-plugin-k8s to deploy your pipelines.

nodestream k8s sync --namespace my-nodestream-app-namespace 

4. Check your kubernetes cluster for your deployed pipelines.

kubectl get cronjobs -n my-nodestream-app-namespace
kubectl get deployments -n my-nodestream-app-namespace

5. Update your nodestream.yaml file and re-run nodestream-plugin-k8s to update your pipelines.

# nodestream.yaml
scopes:
   crons:
      pipelines:
        - path: pipelines/crons/my_scheduled_pipeline.yaml
          annotations:
            nodestream_plugin_k8s_schedule: "30 0 * * *"
    perpetual:
       pipelines:
          - path: pipelines/perpetual/my_kafka_pipeline.yaml
            annotations:
              nodestream_plugin_k8s_conccurency: 100

nodestream k8s sync --namespace my-nodestream-app-namespace 

6. Check your kubernetes cluster for your updated pipelines.

kubectl get cronjobs -n my-nodestream-app-namespace
kubectl get deployments -n my-nodestream-app-namespace

7. Delete your pipelines.

nodestream k8s destroy --namespace my-nodestream-app-namespace 

Annotations

nodestream_plugin_k8s_schedule

This annotation is used to schedule a cronjob. The value of this annotation should be a cron expression.

nodestream_plugin_k8s_concurrency

This annotation is used to set the concurrency of a deployment. The value of this annotation should be an integer.

nodestream_plugin_k8s_image

This annotation is used to set the image of a deployment or job. The value of this annotation should be a valid docker image.