Skip to content

Creating Workflows

Workflows are YAML-defined multi-step pipelines that run as Kubernetes Jobs in the apps namespace.

A workflow is defined in YAML with the following structure:

name: my-workflow
description: Example data processing workflow
steps:
- name: fetch-data
image: python:3.12-slim
command: ["python", "-c", "print('fetching data...')"]
- name: process
image: python:3.12-slim
command: ["python", "/scripts/process.py"]
env:
- name: INPUT_PATH
value: /data/input.csv
  1. Navigate to Workflows
  2. Click New Workflow
  3. Enter the workflow YAML or use the guided form
  4. Click Save

From the Workflows list, click the Run button next to a workflow. A Job resource is created in the apps namespace and the steps execute sequentially.

You can monitor the run status in real time and view logs for each step.

Use environment variables to parameterize your workflows:

steps:
- name: process
image: my-processor:latest
env:
- name: DATASET
value: "{{ .Input.dataset }}"

Pass variables when triggering the workflow from the Run dialog.

Workflows can be scheduled to run automatically using cron syntax:

schedule: "0 2 * * *" # Run daily at 2am

Scheduled workflows appear in the Workflows list with a clock icon.