How do I migrate from Google Workflows to Kestra?

Translate each Workflows step to a Kestra task. HTTP call steps that invoke Cloud Functions can often become Script tasks in Kestra, moving the logic directly into the workflow definition and eliminating the Cloud Function entirely. Steps that call GCP APIs can use Kestra's GCP plugins instead of raw HTTP. Run both platforms in parallel during transition: existing Workflows handle current jobs while new ones go to Kestra.

Can Kestra still use Google Cloud services?

Yes. Kestra has native plugins for BigQuery, Cloud Storage, Pub/Sub, Cloud SQL, Dataflow, and other GCP services. You can also deploy Kestra on GCP using GKE or Cloud Run for the server, and Cloud SQL (PostgreSQL) for state. The difference is that Kestra is not limited to GCP: a single workflow can query BigQuery, write to S3, and trigger a job in Azure without leaving the platform.

Does Kestra run without a GCP account?

Yes. Kestra runs on any infrastructure with Docker or Kubernetes. Two commands and you have a working local instance. For production, deploy to any cloud, your own data center, or Kestra Cloud. If GCP is your primary platform, Kestra deploys there just as easily on GKE or Cloud Run, using Cloud SQL as the state backend.

How does Kestra handle the equivalent of Google Workflows' error handling and retries?

Kestra supports task-level retry policies (fixed, exponential backoff, configurable attempt count), error handlers that run on task failure, and flow-level error tasks for cleanup and alerting. You define these inline in the YAML alongside the task they protect. Google Workflows handles retries through try/except blocks in its DSL syntax; Kestra's approach keeps retry logic declarative and colocated with the task.

What about Google Workflows' pricing model?

Google Workflows charges per workflow step execution: $0.01 per 1,000 steps for internal steps and $0.025 per 1,000 steps for external HTTP calls. For workflows with many steps running frequently, costs accumulate. Kestra open source is free to self-host with no per-step fees. Kestra Enterprise and Kestra Cloud use a subscription model. Calculate your expected step volume against the per-step rate before choosing.

Can Kestra handle the same lightweight service coordination use cases as Google Workflows?

Yes. Kestra's HTTP task calls any external endpoint with configurable auth, headers, body, and retry logic. If your current Google Workflows definition is primarily chaining HTTP calls, the equivalent Kestra flow is straightforward to write. You can also run those HTTP calls alongside direct code execution, database queries, and file operations in the same workflow, without deploying separate Cloud Functions for each piece of logic.

Kestra vs. Google Workflows: Full Orchestration vs. GCP Service Coordination

Google Workflows is a clean, serverless way to coordinate Google Cloud services through HTTP calls. Kestra is what teams reach for when they need to run code directly, orchestrate across clouds, or build data pipelines that go beyond chaining API requests.

Try for Free - 26k GitHub Stars Book a Demo

Two Tools, Two Definitions of "Workflow"

Universal Orchestration: Code, Data, and Events

Open-source platform that runs code directly in isolated containers, orchestrates data pipelines, and reacts to events from any source. Define workflows in declarative YAML, bring scripts in any language, and connect to 1200+ plugins across every major cloud and data tool.

"How do I orchestrate my data pipelines and infrastructure without locking into a single cloud?"

GCP Service Coordination via HTTP

Serverless Google Cloud service for orchestrating HTTP endpoints and GCP APIs. Workflows chains calls to Cloud Functions, Cloud Run services, and GCP APIs using a proprietary YAML/JSON syntax. Code runs in the services being called, not in Workflows itself.

"How do I coordinate my Cloud Functions and GCP services without managing infrastructure?"

GCP Coordination Connects Your Cloud Services.
Workflow Orchestration Automates Your Business.

Business-critical Workflows

Run Python, SQL, Bash, and R directly without wrapping in Cloud Functions
Orchestrate across AWS, Azure, GCP, and on-premises in one platform
1200+ plugins covering databases, data warehouses, messaging, and more
Developer self-service with namespace isolation and RBAC
Event-driven triggers from APIs, message queues, file arrivals, and webhooks

GCP Service Orchestration

Chain HTTP calls to Cloud Functions and Cloud Run services
GCP-native scope only
Proprietary Workflows DSL requires its own learning curve
No direct code execution (logic must live in external services)
Pricing per workflow step execution

Google Workflows is the right choice if you're already running everything on GCP, your logic lives in Cloud Functions or Cloud Run, and you need lightweight HTTP-based coordination without managing any infrastructure. For teams that run code directly, work across clouds, or need a data pipeline tool with a rich plugin ecosystem, Kestra covers more ground without the GCP dependency.

Time to First Workflow

Google Workflows requires a GCP project with billing enabled, enabling the Workflows API, creating a service account with appropriate IAM roles, and deploying the Cloud Functions or Cloud Run services your workflow will call before the workflow itself can do meaningful work.

~5

Minutes

curl -o docker-compose.yml \
https://raw.githubusercontent.com/kestra-io/kestra/develop/docker-compose.yml
docker compose up

# Open localhost:8080
# Pick a Blueprint, run it. Done.

Download the Docker Compose file, spin it up, and you're ready (database and config included). Open the UI, pick a Blueprint, run it. No cloud account required, no IAM to configure.

~30

Minutes

# 1. Enable Workflows API
gcloud services enable workflows.googleapis.com

# 2. Create service account and grant IAM roles
gcloud iam service-accounts create workflows-sa
gcloud projects add-iam-policy-binding $PROJECT_ID \
  --member="serviceAccount:workflows-sa@$PROJECT_ID.iam.gserviceaccount.com" \
  --role="roles/workflows.invoker"

# 3. Deploy your logic as a Cloud Function or Cloud Run service
# 4. Write the workflow YAML and deploy
gcloud workflows deploy my-workflow \
  --source=workflow.yaml \
  --service-account=workflows-sa@$PROJECT_ID.iam.gserviceaccount.com

Requires a GCP project with billing, enabling the Workflows API, creating a service account with IAM bindings, and deploying the Cloud Functions or Cloud Run services your workflow will invoke.

Workflows Your Whole Team Can Read

Kestra: Run Code Directly, No Wrappers Required

id: daily_etl
namespace: analytics

triggers:
  - id: daily
    type: io.kestra.plugin.core.trigger.Schedule
    cron: "0 6 * * *"

tasks:
  - id: run_etl
    type: io.kestra.plugin.scripts.python.Script
    script: |
      import pandas as pd
      from google.cloud import bigquery
      client = bigquery.Client()
      df = client.query("SELECT * FROM dataset.orders").to_dataframe()
      df.to_csv("/tmp/daily_report.csv")

  - id: notify
    type: io.kestra.plugin.notifications.slack.SlackIncomingWebhook
    url: "{{ secret('SLACK_WEBHOOK') }}"
    messageText: "Daily ETL complete"

Write Python, SQL, or Bash inline in your YAML. The AI Copilot generates workflows from a prompt, or start from our library of Blueprints. No Cloud Functions to deploy, no service accounts to wire up per task.

Google Workflows: HTTP Calls to Services That Hold Your Logic

main:
  steps:
    - run_etl:
        call: http.post
        args:
          url: https://us-central1-myproject.cloudfunctions.net/run-etl
          auth:
            type: OIDC
        result: etl_result
    - check_etl:
        switch:
          - condition: ${etl_result.code == 200}
            next: notify
          - condition: true
            next: fail_step
    - notify:
        call: http.post
        args:
          url: https://us-central1-myproject.cloudfunctions.net/notify-slack
          auth:
            type: OIDC
          body:
            message: "Daily ETL complete"
        result: notify_result
    - fail_step:
        raise: ${etl_result}

Google Workflows does not run code. Each task calls an HTTP endpoint (Cloud Function or Cloud Run service) that you deploy and maintain separately. The orchestration YAML wires together those calls; the business logic lives elsewhere.

One Platform for Your Entire Technology Stack

Orchestrate data pipelines, infrastructure operations, business processes, and event-driven workflows across any cloud. Every execution is logged, traceable, and replayable with full visibility into inputs, outputs, and task-level logs.

Designed for chaining GCP services and HTTP endpoints. Works well for lightweight GCP-native automation where logic already lives in Cloud Functions or Cloud Run. Not built for data pipelines, cross-cloud workflows, or executing code outside the GCP HTTP model.

Kestra vs Google Workflows at a Glance


Primary use case	Universal workflow orchestration	GCP service and HTTP endpoint coordination
Code execution	Direct (Python, SQL, Bash, R in containers)	Via HTTP calls to Cloud Functions or Cloud Run only
Workflow definition	Declarative YAML (Git-native)	Proprietary Workflows DSL (YAML/JSON)
Plugin ecosystem	1200+ plugins (databases, warehouses, messaging & more)	GCP APIs and HTTP endpoints only
Cloud support	Multi-cloud (AWS, Azure, GCP, on-premises)	Google Cloud only
Data pipeline support	Native support	Not designed for this
Self-service for non-engineers	Kestra Apps	Not designed for this
Infrastructure automation	Native support	GCP infrastructure only (via API calls)
Open source	Apache 2.0	Proprietary Google Cloud service
Deployment options	Self-hosted, Kestra Cloud, or air-gapped	Google Cloud managed only

Kestra is the unifying layer for our data and workflows. You can start small, but then there is no limit to the possibilities and scalability of such an open architecture.

Julien Henrion, Head of Data Engineering @ Leroy Merlin

+900%In data production

+250Active users

+5,000Workflows created

Kestra Is Built for What Engineering Teams Actually Orchestrate

Code runs here, not somewhere else

Kestra executes Python, SQL, Bash, and R directly in isolated containers. No Cloud Functions to deploy, no Cloud Run services to maintain. Write logic inline in your YAML or reference external scripts; Kestra handles the execution environment.

1200+ plugins, not just HTTP

Connect to BigQuery, Snowflake, dbt, Airbyte, Kafka, S3, PostgreSQL, and 1200+ other tools with purpose-built plugins. Each plugin handles authentication, error handling, and output parsing. An HTTP call can reach any API, but a plugin understands the API.

Not tied to one cloud

Run workflows across AWS, Azure, GCP, and on-premises in a single platform. Deploy Kestra on GCP if that is your primary cloud, and still orchestrate workloads in your data center or another provider. Google Workflows cannot leave GCP.

The Right Tool for the Right Job

Choose Kestra When

You need to run code directly (Python scripts, SQL queries, Bash jobs) without deploying and maintaining Cloud Functions for each task.
Your workflows span more than one cloud, or include on-premises systems that Google Workflows cannot reach.
You're building data pipelines that need native connections to databases, data warehouses, or tools in the modern data stack.
Non-engineers need to trigger and monitor workflows without writing Workflows DSL or managing GCP IAM.
Open source and self-hosted deployment matter for compliance or cost reasons.

Choose Google Workflows When

Everything you orchestrate is already on GCP and lives in Cloud Functions or Cloud Run services.
You need zero infrastructure to manage and are comfortable with per-step GCP pricing.
Your use case is lightweight HTTP-based coordination between GCP services, not data pipeline orchestration.
Deep GCP-native integrations (Cloud Tasks, Pub/Sub, Eventarc) are central to how your system is designed.

Orchestrate Beyond GCP Service Coordination

See how Kestra runs code directly, connects to 1200+ plugins, and works across every cloud — without Cloud Functions.

Book a Demo

Kestra vs. Google Workflows: Full Orchestration vs. GCP Service Coordination

Two Tools, Two Definitions of "Workflow"

Universal Orchestration: Code, Data, and Events

GCP Service Coordination via HTTP

GCP Coordination Connects Your Cloud Services.Workflow Orchestration Automates Your Business.

Business-critical Workflows

GCP Service Orchestration

Time to First Workflow

~5

Minutes

~30

Minutes

Workflows Your Whole Team Can Read

Kestra: Run Code Directly, No Wrappers Required

Google Workflows: HTTP Calls to Services That Hold Your Logic

One Platform for Your Entire Technology Stack

Kestra vs Google Workflows at a Glance

Kestra Is Built for What Engineering Teams Actually Orchestrate

Code runs here, not somewhere else

1200+ plugins, not just HTTP

Not tied to one cloud

The Right Tool for the Right Job

Choose Kestra When

Choose Google Workflows When

Frequently asked questions

How do I migrate from Google Workflows to Kestra?

Can Kestra still use Google Cloud services?

Does Kestra run without a GCP account?

How does Kestra handle the equivalent of Google Workflows' error handling and retries?

What about Google Workflows' pricing model?

Can Kestra handle the same lightweight service coordination use cases as Google Workflows?

Orchestrate Beyond GCP Service Coordination

GCP Coordination Connects Your Cloud Services.
Workflow Orchestration Automates Your Business.