Deep Dive into the New Features of Argo Workflows 3.6

Argo Workflows, a CNCF graduated project, is designed to orchestrate parallel jobs on Kubernetes; this article examines the key new features introduced in version 3.6.

The platform is widely used for batch processing, machine‑learning pipelines, infrastructure automation, and CI/CD, offering a declarative way to define and run large‑scale tasks.

Alibaba Cloud Container Service, an early adopter, contributed numerous enhancements to v3.6, including massive‑parameter offload, template scheduling constraints, large‑scale flat workflow parallel parsing, OSS artifact streaming and garbage collection, pod parallel cleanup with retry, and dynamic template references, all aimed at improving stability, usability, and performance.

1) CronWorkflows – scheduling enhancements – support for multiple cron schedules in a single workflow, a stop‑strategy to prevent backlog of failing runs, and a when expression for conditional execution. Example:

apiVersion: argoproj.io/v1alpha1
kind: CronWorkflow
metadata:
  name: cron-workflow-example
spec:
  schedules:
  - "*/3 * * * *"
  - "*/5 * * * *"
  concurrencyPolicy: "Allow"
  stopStrategy:
    condition: "failed >= 10"
  when: "{{= cronworkflow.lastScheduledTime == nil || (now() - cronworkflow.lastScheduledTime).Seconds() > 3600 }}"
  startingDeadlineSeconds: 0
  workflowSpec:
    entrypoint: whalesay
    templates:
    - name: whalesay
      container:
        image: alpine:3.6
        command: [sh, -c]
        args: ["date; sleep 1"]

2) UI improvements – added detailed workflow information, time stamps, artifact directory view, Markdown support, and access to execution history and real‑time logs for both CronWorkflows and WorkflowTemplates.

3) Controller upgrades – enhancements for large‑scale stability and security, including queued workflow archiving, parallel pod cleanup, finalizers to avoid premature pod deletion, massive flat‑workflow parallel parsing, automatic seccomp profile setting, and more.

a) OSS artifact automatic garbage collection with configurable artifactGC strategy. Example:

apiVersion: argoproj.io/v1alpha1
kind: Workflow
metadata:
  generateName: artifact-gc-
spec:
  entrypoint: main
  artifactGC:
    strategy: OnWorkflowDeletion
  templates:
  - name: main
    container:
      image: argoproj/argosay:v2
      command: [sh, -c]
      args: ["echo \"hello world\" > /tmp/on-completion.txt"]
    outputs:
      artifacts:
      - name: on-completion
        path: /tmp/on-completion.txt
        oss:
          endpoint: http://oss-cn-zhangjiakou-internal.aliyuncs.com
          bucket: my-argo-workflow
          key: on-completion.txt
          accessKeySecret:
            name: my-argo-workflow-credentials
            key: accessKey
          secretKeySecret:
            name: my-argo-workflow-credentials
            key: secretKey
        artifactGC:
          strategy: OnWorkflowCompletion

b) Template scheduling constraints via nodeSelector and tolerations . Example:

apiVersion: argoproj.io/v1alpha1
kind: WorkflowTemplate
metadata:
  name: benchmarks
spec:
  entrypoint: main
  serviceAccountName: workflow
  templates:
  - dag:
      tasks:
      - name: benchmark
        template: benchmark
  - name: main
    nodeSelector:
      pool: workflows
    tolerations:
    - key: pool
      operator: Equal
      value: workflows

c) Dynamic template references allowing parameters to be injected directly into templateRef fields, reducing YAML size and improving readability. Example:

apiVersion: argoproj.io/v1alpha1
kind: Workflow
metadata:
  generateName: hello-world-wf-global-arg-
  namespace: default
spec:
  entrypoint: whalesay
  arguments:
    parameters:
    - name: global-parameter
      value: hello
  templates:
  - name: whalesay
    steps:
    - - name: hello-world
        templateRef:
          name: '{{item.workflow-template}}'
          template: '{{item.template-name}}'
        withItems:
        - {workflow-template: 'hello-world-template-global-arg', template-name: 'hello-world'}
    - name: hello-world-dag
      template: diamond
  - name: diamond
    dag:
      tasks:
      - name: A
        templateRef:
          name: '{{item.workflow-template}}'
          template: '{{item.template-name}}'
        withItems:
        - {workflow-template: 'hello-world-template-global-arg', template-name: 'hello-world'}

d) Updated expression library supporting additional functions such as list concatenation and string joining. Example:

apiVersion: argoproj.io/v1alpha1
kind: Workflow
metadata:
  generateName: test-expression-
  namespace: argo
spec:
  entrypoint: main
  arguments:
    parameters:
    - name: expr
      value: "{{= concat(['a', 'b'], ['c', 'd']) | join('\n') }}"
  templates:
  - name: main
    inputs:
      parameters:
      - name: expr
    script:
      image: alpine:3.6
      command: ["sh"]
      source: |
        echo result: '{{ inputs.parameters.expr }}'"

4) Argo CLI enhancements – simplified commands for updating CronWorkflows, workflow templates, and cluster‑templates, and label‑based filtering for template management, e.g., argo template list -l app=test .

Finally, the article highlights that Argo Workflows is a cloud‑native batch‑task engine suitable for architects, data scientists, and DevOps engineers, and mentions Alibaba Cloud’s fully managed Serverless Argo Workflows offering with simple REST API and Python SDK, high performance, and extensive best‑practice documentation.