What’s New in Argo Workflows at KubeCon Europe 2025? Highlights and Insights

Community Growth

Argo Workflows’ GitHub repository grew from 14.0k to 15.5k stars. Five new maintainers from four continents joined, including two end‑users, reflecting a diversified contributor base. The 2024 focus was on improving user experience and orchestration performance, with a roadmap that adds enterprise‑grade scaling, AI/Big Data integration, tracing, and security features.

Unified Orchestration of AI/HPC Workloads

Sebastian Beyvers (Giessen University) described a migration strategy that consolidates Kubeflow, Airflow, and Slurm workloads onto a single Argo Workflows platform. The method uses reusable WorkflowTemplate objects, wraps external tools in container or script steps, and leverages parameterization to run the same workflow across multiple environments, delivering a unified architecture and a smoother developer experience.

Event‑Driven Workflows with Argo Events

Darko Janjić (Pipekit) showed how to build event‑driven pipelines using Argo Events. Key techniques include custom event filters in Sensor specifications, integration with CI/CD pipelines, security scanners, and automated incident remediation. An example sensor definition is shown below:

apiVersion: argoproj.io/v1alpha1
kind: Sensor
metadata:
  name: github-push-sensor
spec:
  dependencies:
  - name: github-push
    eventSourceName: github-source
    eventName: push
  triggers:
  - template:
      name: build-pipeline
      k8s:
        group: argoproj.io
        version: v1alpha1
        resource: workflows
        operation: create
        source:
          inline:
            workflow:
              spec:
                entrypoint: main
                templates:
                - name: main
                  container:
                    image: alpine
                    command: [sh, -c, "echo Build triggered"]

Platform Engineering Comparison

Viktor Farcic (Upbound) and Mauricio Salatino (Diagrid) compared Argo Workflows with Tekton Pipelines. Tekton excels at pure CI/CD pipelines, while Argo Workflows provides broader capabilities for parallel data processing and MLOps. Demonstrations highlighted Argo’s DAG and step templates for complex dependency graphs, enabling multi‑stage ML pipelines that go beyond simple build‑test‑deploy cycles.

Real‑World Use Cases

Anjelica Ambrosio (Akuity) explained how to scale time‑series analysis by using withParam to launch parallel processing pods, achieving high‑throughput data ingestion.

Will Wang (Bloomberg) presented the use of validating webhooks to enforce workflow spec schemas, reducing runtime failures in large‑scale ML pipelines.

Chris Heisz & Luke Beamish (Neo4j) demonstrated linking Argo Workflows with Backstage to automate chaos testing, triggering fault‑injection experiments as discrete workflow steps.

References

GitHub repository: https://github.com/argoproj/argo-workflows

One Engine To Rule Them All – https://colocatedeventseu2025.sched.com/event/1u5dO/one-engine-to-rule-them-all-unifying-cloud-workloads-with-argo-workflows-sebastian-beyvers-giessen-university

Automated Resilience with Argo Events – https://colocatedeventseu2025.sched.com/event/1u5dC/automated-resilience-using-argo-events-for-real-time-incident-remediation-darko-janjic-pipekit

The Past, Present, and Future of Platform Engineering – https://colocatedeventseu2025.sched.com/event/1u5lu/the-past-the-present-and-the-future-of-platform-engineering-mauricio-salatino-diagrid-viktor-farcic-upbound

Scaling Time Series Analysis – https://www.youtube.com/watch?v=jXbapwPfB7Q&list=PLj6h78yzYM2N9MWCsU_4upn64NDtHGv6i&index=26

Spec Validation Webhooks – https://www.youtube.com/watch?v=h-WHhA-T4lE&list=PLj6h78yzYM2N9MWCsU_4upn64NDtHGv6i&index=27

From Click To Chaos – https://colocatedeventseu2025.sched.com/event/1u5en/from-click-to-chaos-linking-argo-workflows-and-backstage-for-automated-testing-chris-heisz-luke-beamish-neo4j