Jump to

Cloud Deploy

Deliver continuously to Google Kubernetes Engine, Cloud Run, and more.

  • Create deployment pipelines for GKE, Cloud Run, and more

  • Fully managed continuous delivery service for easy scaling

  • Enterprise security and audit

  • Built-in delivery metrics

  • Snaps into your existing DevOps ecosystem

Cloud Deploy icon

Key features

Key features

Streamlined continuous delivery

Cloud Deploy makes continuous delivery to GKE, Cloud Run services and jobs, and Anthos easy and powerful. Define releases and progress them through environments, such as test, stage, and production. Cloud Deploy provides easy one-step promotion and rollback of releases via the web console, CLI, or API. Built-in metrics enable insight into deployment frequency and success.

Fully managed single pane of glass

As a fully managed service, Cloud Deploy has no infrastructure to set up and manage, while providing scale-up and scale-down automatically to optimize cost and performance. This centralization also provides a single pane of glass to monitor and control release candidates organization-wide as they progress toward production.

Tightly integrated with Google Cloud

Cloud Deploy is the most integrated GKE, Cloud Run, and Anthos deployment platform available. Lockdown release progression via IAM, monitor release events with Cloud Logging, and achieve traceability with Cloud Audit Logs. Connect monitoring to deployed resources.

Integrates with the tools you love

Cloud Deploy can be integrated with popular DevOps tools, such as CI and ticketing. Cloud Deploy brings Skaffold to your pipelines, which, in unison with Cloud Code, brings pipeline parity across dev and CI/CD.

View all features

"Cloud Deploy offers a streamlined approach to create CI/CD pipelines using Skaffold, along with advanced features like canary deployment and verification. Additionally, it offers a unified developers’ experience for GKE and Cloud Run, making it easy to choose the most suitable platform for applications."

Jun Sakata - Head of Platform Engineering, Ubie



Google Cloud Basics
Cloud Deploy concepts

Learn more about Cloud Deploy, including how it works, terminology, and architecture.

Deploy an application to two GKE targets

Deploy a simple app to a progression of Google Kubernetes Engine clusters.

Deploy an application to two Cloud Run targets

Deploy a simple app to a progression of Cloud Run services.

Cloud Deploy tutorials

These hands-on tutorials guide you through setting up a pipeline and deploying a sample application using Cloud Deploy.

All features

Learn more about Cloud Deploy features

Pipeline visualization Visualize the path to delivery. Define delivery pipelines and visualize the progression of candidate releases through to production.
Easy rollout/rollback Rollout and rollback to GKE, Cloud Run, and Anthos user clusters has never been easier and more clear. Promote a release between target stages using a one-step operation in the web console, CLI, or API.
Built-in approvals Cloud Deploy supports separation of duties and concerns with formal release promotion approvals, accessible via the web console, CLI, or API and integrated with IAM.
Parallel deploy Deploy to multiple GKE or Anthos clusters, or Cloud Run service regions, concurrently. Cloud Deploy orchestrates to ensure the deployment succeeds everywhere or is collectively rolled back.
Canary deploy Progressively deploy a new version of your application to a specified portion (such as 10%) of traffic. 
Deploy hooks Configure Cloud Deploy to perform pre-deployment, post-deployment actions, or both.
Deployment verification Integrate deployment and verification tests to have Cloud Deploy confirm rollout success. 
Automation Configure continuous deployment within your delivery pipeline through automation. Automatically promote releases from one target to the next as well as automate rollout canary percentages. 
Declarative configuration Never worry about the how, just define the what and let Cloud Deploy do the heavy lifting. Cloud Deploy fully manages GKE, Cloud Run, and Anthos user cluster deployments based on desired end states.
Custom target types Custom target types extend Cloud Deploy by allowing you to define and use a custom target type with its own renderer and deployer while continuing to make use of Cloud Deploy features, including approval and promotion.
Opinionation Cloud Deploy provides an opinionated on-ramp to GKE, Cloud Run, and Anthos via Skaffold. Leverage built-in best practices, which keep pipelines durable by insulating them against changes.
Tightly integrated No hand wiring required. Cloud Deploy comes pre-integrated to IAM, Cloud Logging, and Cloud Audit Logs.
Metrics Insights at your fingertips. Know how frequently and successfully releases progress through delivery pipelines.
Auditing and traceability Cloud Deploy integrates with Cloud Logging to provide release auditability and traceability. Maintain clarity on which releases were promoted and by whom.
IAM and execution permissioning Lockdown release deployments with granular IAM permissioning and scoped service accounts for execution.
Connect the tools you love Cloud Deploy extends your DevOps ecosystem and plays with all the tools you love. Invoke Cloud Deploy from popular CI products using the CLI or API and federate approvals to ticketing systems of choice.
Unified with your developer experience Cloud Deploy brings Skaffold to your pipelines, enabling operators to achieve pipeline parity across dev and CI/CD, while developers remain productive and insulated from platform changes while leveraging the idiomatic developer experience of Cloud Code.



Cloud Deploy customers are charged a management fee per active delivery pipeline with more than one target (“multiple target delivery pipeline”). There is no charge for the first active multiple target delivery pipeline per billing account each month, with each additional active multiple target delivery pipeline charged at $5 per month. 

Delivery pipelines with only one target do not incur a management fee.

For all delivery pipelines, charges for underlying services also apply.