Google Cloud Workflows
cloud.google.com
- Free plan
- Web
- API
What Google Cloud Workflows is
Google Cloud Workflows is a managed platform for coordinating services in a defined order, including Google Cloud services, custom services, and HTTP APIs. Workflows are written in YAML or JSON and can use conditions, loops, parallel steps, and reusable subworkflows. Executions can start manually, through client libraries or REST, or on a schedule with Cloud Scheduler. Connectors link workflows to Google Cloud APIs such as Pub/Sub, Firestore, and Secret Manager, while workflow steps can also make HTTP or HTTPS requests to custom and private endpoints. Reliability features include retries, polling, error handling, callbacks, and waits lasting up to one year. Access is controlled with IAM, and Secret Manager can store and retrieve secrets. Cloud Logging and Cloud Monitoring provide execution logs and visibility into volume, errors, and execution time. Pricing is based on executed steps, including failed and retried steps; idle workflows incur no charges. An execution can run for up to one year and perform at most 100,000 steps. New Google Cloud customers get $300 in free credits to spend on Workflows.
Who it suits
Workflows suits teams coordinating Google Cloud services, custom services, or APIs, including for batch jobs and business or IT process automation. It fits users comfortable defining workflows in YAML or JSON rather than using a visual builder.
Good to know, and what to watch
- Supports conditions, loops, and parallel steps.
- Connectors handle retries and long-running operations.
- Integrates with Cloud Logging and Cloud Monitoring.
- Idle workflows do not accrue charges.
- Executions are limited to 100,000 steps.
- Maximum execution duration is one year.
- Pricing includes failed and retried steps.
- No visual workflow builder is listed.
Plans
Every plan and priceAlways Free
Always Free external steps
Always Free internal steps
Internal steps
Free plan
Getting started
- Open the Google Cloud Workflows website
- Define the workflow in YAML or JSON
- Choose Google Cloud connectors or HTTP and HTTPS steps
- Set a manual, programmatic, or Cloud Scheduler trigger
- Assign IAM permissions and use Secret Manager for secrets when needed
- Start with the free plan or trial, then use pay-as-you-go billing from $0.01
Limits to check
The Always Free plan includes 5,000 internal steps and 2,000 external steps monthly. A workflow execution is limited to 100,000 steps and one year; pay-as-you-go charges apply to steps beyond the free allowances, including failed and retried steps.
The Club's review
Google Cloud Workflows is a managed service for coordinating cloud services and API calls as defined processes. It suits teams comfortable working in YAML or JSON, particularly those using Google Cloud. Its connectors and long-running executions are useful strengths, though per-step billing and an execution cap shape what it can handle well.
Overview
Workflows runs sequences that can include Google Cloud services, custom services, and HTTP APIs. Definitions can combine conditions, loops, parallel steps, and reusable subworkflows, making the service suitable for more than linear handoffs. Executions can be started manually, through client libraries or REST, or on a schedule with Cloud Scheduler. Google identifies service orchestration, batch jobs, business process automation, and IT process automation as use cases.
The cloud deployment removes the need to operate an orchestration service yourself, but it is not a visual process designer: workflows are written in YAML or JSON, and there is no visual builder. That makes it a better fit for teams comfortable treating workflow logic as code than for people seeking a canvas-based tool.
Key features
Connectors and API calls
Connectors integrate with Google Cloud APIs such as Pub/Sub, Firestore, and Secret Manager. They handle request formatting, retries, and long-running operations, reducing integration work for processes centered on Google Cloud. Workflow steps can also call custom or private endpoints over HTTP or HTTPS, so the service is not limited to Google's own services.
Reliability and execution control
Retries, polling, error handling, and callbacks help manage failure and asynchronous responses. A workflow can wait for up to one year, which supports processes with long pauses rather than only short-lived jobs. The trade-off is a maximum of 100,000 steps per execution; workloads that exceed that ceiling need to be redesigned or run elsewhere.
Security and observability
Workflows runs in a sandbox without code or library dependencies that need security patches. It can retrieve secrets from Secret Manager. IAM governs access, and connectors use built-in IAM authentication, but the workflow service account still needs the permissions required for its calls. Cloud Logging and Cloud Monitoring provide execution logs and visibility into volume, errors, and execution time, useful for operators diagnosing failures or tracking throughput.
Pricing
The pricing model is freemium, with an Always Free plan at no cost, billed Monthly free usage limits. It includes 5,000 internal steps and 2,000 external steps. That is a practical starting point for small workloads, but the caps matter once executions become frequent or involve many API calls.
Pay as you go costs $0.01 a month, billed per increment of 1,000 steps per month for internal steps after the first 5,000. External steps after the first 2,000 cost $0.025 per increment of 1,000. Pricing is based on executed steps, including failed and retried steps; idle workflows incur no charges. After 100 million internal or external steps, contact sales. New Google Cloud customers receive $300 in free credits to spend on Workflows, and Google Cloud customers, including free trial customers, receive Basic Support with documentation, community forums, billing assistance, and Active Assist.
Step-based billing makes the free allowance and retry behavior important when estimating costs: failed or retried work counts, while waiting idle does not. The stated paid rate is per increment, not a flat rate for unlimited use, so high-volume teams should account for step counts and the sales threshold.
Platforms
Workflows is available through the web and API and is deployed in the cloud. It supports event triggers and automatic retries, with execution observability. It does not provide human approvals, so processes requiring an approval step need another way to handle that stage.
Who it's for
Workflows is a strong fit for developers and cloud operations teams that want managed orchestration, especially when their processes use Google Cloud APIs and need retries, monitoring, or long waits. It also accommodates custom and private HTTP endpoints. It is a weaker choice for teams that need a visual builder, human approval stages, self-hosting, or executions beyond 100,000 steps.
Pros and cons
- Pros: Google Cloud connectors handle request formatting, retries, and long-running operations, reducing integration work for cloud-centric flows.
- Pros: Executions can wait up to a year, with retries, callbacks, and monitoring for long-running processes.
- Pros: Idle workflows do not accrue charges, and the free tier provides monthly internal and external step allowances.
- Cons: YAML or JSON definitions and the absence of a visual builder make it less suitable for canvas-first workflow design.
- Cons: Failed and retried steps count toward charges, so recovery behavior affects cost.
- Cons: The 100,000-step execution limit and lack of human approvals constrain some larger or approval-driven processes.
Alternatives
For a free, Apache 2.0-licensed platform, consider Kestra. Windmill is worth considering when a free and open-source plan with stated workspace, user, storage, and guest-app allowances suits the team. Choose Argo Workflows for an open-source workflow engine for Kubernetes with no paid plans listed. Azure Logic Apps is an alternative with Consumption and Standard plans, including a Consumption allowance of 4,000 actions.
CloudBolt offers a free CMP plan with core capabilities for up to 100 resources. HCL Workload Automation is an enterprise workload automation alternative with a paid plan. Juju is a free, open-source orchestration engine. Temporal offers a usage-based Pay-As-You-Go plan at $50 a month, with no base monthly fee and $150 in free credits for 90 days.
Browse IT Orchestration Software, Process Orchestration Software, and Workflow Orchestration Software for more options.
Verdict
Choose Google Cloud Workflows if you want managed, code-defined orchestration and your flows benefit from Google Cloud connectors, retries, and execution monitoring. Its cloud operation and long wait allowance are useful for persistent processes. Look elsewhere if you need a visual designer, human approvals, self-hosted deployment, or executions that can exceed 100,000 steps; budget carefully because retries and failures count as billable steps.
The verdict
Workflows offers code-defined orchestration with connectors, retries, and execution monitoring. Consider its per-step pricing and execution limits when planning workloads.
Questions people ask
Does Google Cloud Workflows offer a free plan or trial?
Yes. It has an Always Free plan and a free trial. New Google Cloud customers also receive $300 in credits to spend on Workflows.
How is Workflows priced?
The listed paid plan starts at $0.01 a month per increment of 1,000 steps. Charges are based on executed steps, including failed and retried steps; idle workflows do not accrue charges.
Which services can Workflows connect to?
Connectors include Google Cloud APIs such as Pub/Sub, Firestore, and Secret Manager. Workflow steps can also call custom and private HTTP or HTTPS endpoints.
Can executions run on a schedule?
Yes. Executions can start manually, through client libraries or REST, or on a schedule using Cloud Scheduler.
Is there a visual workflow builder?
No. Workflow definitions are written in YAML or JSON.
What the maker publishes
- Access control
- Workflows uses IAM, and connectors use built-in IAM authentication while requiring the workflow service account to have the needed permissions. source
- API surface
- The documentation provides Workflows and Workflow Executions REST APIs, gcloud CLI commands, syntax and connector references, and client-library references. source
- Authentication
- Connectors use IAM-based service-account authentication, but the workflow service account still needs permissions for each target API. source
- Breaking change
- From May 2, 2025, Workflows no longer supports HTTP requests to IP-based GKE control-plane endpoints; DNS endpoints are required. source
- Callbacks
- A workflow can create a callback URL and wait for external or human-in-the-loop input for up to one year. source
- Company founding
- Google's company history says Google Inc. was officially established in August 1998. source
- Company ownership
- Google is a subsidiary of Alphabet, according to its Crunchbase organization profile. source
- Company profile
- LinkedIn lists Google as a public software-development company headquartered in Mountain View, California, with 10,001+ employees. source
- Compliance
- Google lists Workflows as compliant with C5:2020, FedRAMP High, HIPAA, HITRUST CSF, ISO 27001/27017/27018, PCI DSS, and SOC 1/2/3, excluding Preview features. source
- Connector integrations
- Documented connector examples include Pub/Sub, Firestore, Secret Manager, BigQuery, Google Sheets, Cloud Translation, Cloud Run, Batch, and Kubernetes APIs. source
- Connectors
- Connectors simplify Google Cloud API calls by formatting requests and providing built-in retries and long-running-operation handling. source
- Documentation depth
- The documentation set includes quickstarts, tutorials, syntax and standard-library references, REST APIs, connectors, quotas, release notes, training, and code samples. source
- Encryption
- Customer-managed Cloud KMS keys can encrypt workflow definitions and execution data at rest. source
- Execution limits
- A workflow execution can contain up to 100,000 steps and run for at most one year. source
- External APIs
- Workflow steps can make HTTP or HTTPS requests to APIs, including custom and private endpoints. source
- Formats
- Workflow definitions use YAML or JSON, and Google publishes JSON-schema support for syntax autocompletion. source
- Free credits
- New Google Cloud customers get $300 in free credits to spend on Workflows. source
- Infrastructure as code
- Google documents creating, deploying, and executing Workflows with Terraform and the Google provider. source
- Integrations
- Connectors integrate workflows with Google Cloud APIs and handle request formatting, retries, and long-running operations; examples include Pub/Sub, Firestore, and Secret Manager. source
- Licence
- The managed Workflows service is proprietary; Google's public Workflows samples repository is marked Apache-2.0. source
- Limit: learning curve
- A G2 reviewer says the product can be complex initially and requires a solid understanding to use effectively. source
- Limit: pipeline controls
- A G2 reviewer requests additional pipeline-control capabilities. source
- Limit: pricing
- G2 reviewers complain that pricing can become costly for complex flows and that the free quota is small. source
- Monitoring
- Workflows integrates with Cloud Logging and Cloud Monitoring for execution logs, volume, error rates, and execution time. source
Checked against the maker's pages on 15 Oct 2026.