Workflows / Cloud Tasks Complete Guide: Serverless Orchestration

Workflows and Cloud Tasks are GCP's serverless orchestration services. Workflows runs APIs/services in sequence; Cloud Tasks is an async task queue. Using them together lets you build complex serverless flows.

Workflows

Key Features

• Serverless, YAML / JSON based
• Integrates with HTTP / GCP APIs / Cloud Functions / Cloud Run
• Control structures: parallel, for loops, try-except, switch
• Maximum execution time: 1 year
• Connectors give one-click access to 50+ GCP services
• Event-driven via Eventarc integration

Workflow YAML Example

main:
  params: [input]
  steps:
    - validate:
        call: http.post
        args:
          url: https://api.example.com/validate
          body: ${input}
        result: validation
    - check:
        switch:
          - condition: ${validation.body.valid}
            next: process
          - condition: true
            return: "Invalid input"
    - process:
        parallel:
          shared: [result]
          branches:
          - charge:
              steps:
              - call_charge:
                  call: googleapis.cloudfunctions.v2.functions.call
                  args:
                    name: projects/PROJECT/locations/asia-northeast1/functions/charge
                    body: ${input}
                  result: result.charge
          - notify:
              steps:
              - call_notify:
                  call: googleapis.cloudfunctions.v2.functions.call
                  args:
                    name: projects/PROJECT/locations/asia-northeast1/functions/notify
                    body: ${input}
                  result: result.notify
    - return:
        return: ${result}

Cloud Tasks

Key Features

• HTTP targets and App Engine targets
• Delayed execution (scheduleTime)
• Rate limiting (Token Bucket / Concurrent)
• Retry configuration (max attempts / min backoff / max backoff)
• Dead Letter Queue support
• Automatic OIDC / OAuth token injection

Task Creation Example (Python)

from google.cloud import tasks_v2

client = tasks_v2.CloudTasksClient()
parent = client.queue_path('PROJECT', 'asia-northeast1', 'my-queue')

task = {
    "http_request": {
        "http_method": tasks_v2.HttpMethod.POST,
        "url": "https://my-service.run.app/process",
        "body": json.dumps({"user_id": 123}).encode(),
        "headers": {"Content-Type": "application/json"},
        "oidc_token": {
            "service_account_email": "[email protected]",
        },
    },
    "schedule_time": datetime.utcnow() + timedelta(minutes=10),
}
response = client.create_task(parent=parent, task=task)

Pricing

When to Use Which

Typical Use Cases

Workflows

• Customer onboarding (Auth → DB registration → email → Slack notification)
• Vertex AI inference pipeline (image fetch → preprocessing → inference → result storage)
• Multi-cloud integration (AWS S3 → Dataflow → BQ)

Cloud Tasks

• Delayed user notifications (24 hours after sign-up)
• Rate-limiting webhook calls
• Async report generation
• Reliable calls to external APIs (retry + backoff)

Comparison with Other Clouds

Google Cloud is a trademark of Google LLC. For the latest information, see the official Workflows documentation.