Cloud Run

Cloud Run is a fully managed serverless compute platform from Google Cloud, a division of Google LLC (Alphabet Inc.). It lets developers deploy stateless containers that scale automatically — including down to zero — and are billed only for the resources consumed while requests are being handled. Because it is built on the open-source Knative standard and runs on Google's managed infrastructure, Cloud Run abstracts away cluster provisioning, patching, and capacity planning, positioning it between fully managed function services such as Cloud Functions and fully configurable container orchestration such as Google Kubernetes Engine (GKE). The platform is used for web applications and APIs, microservices, event-driven workloads, batch jobs, and internal tools. It integrates with the broader Google Cloud stack, including Cloud Build, Artifact Registry, Cloud SQL, Pub/Sub, Cloud Storage, and Identity and Access Management, and it supports custom domains, HTTPS, traffic splitting for canary releases, and concurrency controls. Cloud Run also offers a jobs execution model for run-to-completion workloads and a functions-style deployment experience for developers who prefer source-based workflows. Cloud Run is aimed at developers, platform teams, and enterprises that want container portability without operating Kubernetes. It competes with AWS Fargate and AWS App Runner, Microsoft Azure Container Apps, and other serverless container offerings. Google positions it as a way to modernize applications, reduce operational overhead, and pay only for what is used, with regional availability across many Google Cloud locations worldwide. As part of Google Cloud, Cloud Run benefits from Google's global network, security, and compliance programs. It is generally available and continues to receive new features, including GPU support for AI inference and other workload types, reflecting Google Cloud's broader push into AI and serverless computing.

Cloud Run was introduced by Google Cloud as a managed compute platform for containerized applications, building on the open-source Knative project that Google helped create to bring serverless primitives to Kubernetes. It emerged as part of Google's effort to offer a continuum of compute options: Cloud Functions for event-driven functions, App Engine for opinionated application hosting, GKE for full Kubernetes control, and Cloud Run for teams that want containers without cluster management. The service became generally available and was positioned as a serverless container runtime that could scale to zero and bill only for actual usage. Over time Google expanded the platform with features such as Cloud Run jobs for batch and run-to-completion workloads, traffic splitting for gradual rollouts and canary deployments, custom domains and managed TLS, concurrency settings, and integrations with Cloud Build, Artifact Registry, and other Google Cloud services. Google also introduced a functions-style experience on Cloud Run, allowing developers to deploy source code without writing a Dockerfile. As AI workloads grew, Google added GPU support to Cloud Run so that inference and other accelerated workloads could run on the same serverless model. Cloud Run is available in many Google Cloud regions worldwide and is marketed alongside GKE Autopilot and other managed compute services. Its development reflects the broader industry shift toward serverless containers and the convergence of functions, containers, and Kubernetes-based platforms.

• Google Cloud introduces Cloud Run as a managed serverless container platform built on Knative.

• Cloud Run becomes generally available with scale-to-zero autoscaling and pay-per-use billing.

• Cloud Run jobs are added for batch and run-to-completion workloads.

• A functions-style deployment experience is introduced on Cloud Run for source-based workflows.

• GPU support is added to Cloud Run for AI inference and accelerated workloads.

Developer-friendly, fully managed serverless container platform that bridges functions and Kubernetes, with pay-per-use pricing and open-standard portability.

• Cloud Run (services)

  Fully managed serverless container runtime for request-driven applications and APIs.

• Cloud Run jobs

  Execution model for batch and run-to-completion containerized workloads.

• Cloud Run functions

  Source-based deployment experience for functions on the Cloud Run platform.

1. 1Cloud Run official product pagecloud.google.com
2. 2Cloud Run documentationcloud.google.com
3. 3Google Cloud Introduces Option to Choose Region with Lowest CO2 Footprintinfoq.com
4. 4Google unveils Cloud Run instances for always-on AI agentsitbrief.news