Google Cloud Run

Google Cloud Run is a managed compute platform that runs stateless containers and scales them automatically, including down to zero when idle. It is part of Google Cloud and bills on a usage-based, pay-per-use model.

Visit Google Cloud Run →

Some links may be affiliate links — we may earn a commission if you sign up, at no extra cost to you. It never affects our rankings. Learn more.

Google Cloud Run is a fully managed service for running containerized applications without provisioning or managing servers. It accepts any container image that listens on a port, automatically scaling instances up in response to incoming requests or events and down to zero when there is no traffic. Because it is built on the open-source Knative API and runs standard containers, workloads remain portable across environments. Cloud Run integrates with the wider Google Cloud ecosystem, including Cloud Build for source-based deployments, Artifact Registry, IAM, and VPC networking. Billing is usage-based, charging for CPU, memory, and requests only while instances are active. It supports both request-driven services and background jobs.

Key features

Runs any standard container image (Docker/OCI) that serves on a port
Automatic scaling from zero to many instances based on load
Usage-based billing for CPU, memory, and requests with per-100ms granularity
Continuous deployment from source via Cloud Build or Developer Connect (GitHub, GitLab, Bitbucket)
Built on the open Knative API for container portability
Integration with GCP IAM, VPC, Secret Manager, and Cloud Run Jobs for batch tasks

Specs

Category	Container hosting
Pricing model	usage-based
Free tier	Yes — Perpetual monthly free allowance: 2M requests, 180,000 vCPU-seconds, 360,000 GiB-seconds of memory, and 1 GiB of North America egress.
Managed database	No
Regions	40+ (GCP regions)
Self-hostable	No
Docker-native	Yes
Git deploy	Yes
Open source	No

Last reviewed 2026-07-16 — see our methodology. Specs, pricing and free tiers change often, so confirm current details on the provider before deciding.

Pros & cons

Our editorial assessment — not vendor-stated facts.

Pros

Scale-to-zero means no charge when the service is idle
Deploys any container image without framework or language lock-in
Generous perpetual free monthly allowance for low-traffic workloads
Deep integration with the broader Google Cloud platform and services
Automatic HTTPS, request-based autoscaling, and managed infrastructure

Cons

Cold starts can add latency when scaling from zero unless minimum instances are set
GCP's IAM, networking, and project setup add a learning curve versus simpler PaaS tools
Network egress and cross-service traffic can add costs beyond compute
Per-request/usage billing makes costs harder to predict for sustained high traffic
No built-in managed database; state must come from separate GCP services

Pricing

Cloud Run uses a pay-per-use model, billing for allocated CPU and memory (per vCPU-second and GiB-second) plus a per-request charge, generally only while instances are handling requests. A perpetual free tier covers roughly 2 million requests, 180,000 vCPU-seconds, 360,000 GiB-seconds of memory, and 1 GiB of North America egress per month. Because services scale to zero, an idle deployment incurs no compute charges, though egress and dependent GCP services are billed separately.

Who should use Google Cloud Run?

Cloud Run suits teams already on Google Cloud who want to run containers without managing servers, especially for spiky, event-driven, or request-based workloads. It is a strong fit when scale-to-zero economics matter and when integration with GCP services like Cloud Build, Pub/Sub, or IAM is valuable. Teams wanting the simplest possible deploy experience without cloud-platform complexity may prefer a lighter PaaS.

Frequently asked questions

Does Google Cloud Run have a free tier?: Yes. Perpetual monthly free allowance: 2M requests, 180,000 vCPU-seconds, 360,000 GiB-seconds of memory, and 1 GiB of North America egress.
How much does Google Cloud Run cost?: Cloud Run uses a pay-per-use model, billing for allocated CPU and memory (per vCPU-second and GiB-second) plus a per-request charge, generally only while instances are handling requests. A perpetual free tier covers roughly 2 million requests, 180,000 vCPU-seconds, 360,000 GiB-seconds of memory, and 1 GiB of North America egress per month. Because services scale to zero, an idle deployment incurs no compute charges, though egress and dependent GCP services are billed separately.
Who should use Google Cloud Run?: Cloud Run suits teams already on Google Cloud who want to run containers without managing servers, especially for spiky, event-driven, or request-based workloads. It is a strong fit when scale-to-zero economics matter and when integration with GCP services like Cloud Build, Pub/Sub, or IAM is valuable. Teams wanting the simplest possible deploy experience without cloud-platform complexity may prefer a lighter PaaS.