containerized serverless

Google Cloud Run

Google Cloud Run is a managed serverless platform that deploys and runs scalable containerized applications directly on Google infrastructure.

What the product actually does

Cloud Run accepts any standard OCI container image and deploys it as a scalable web service or batch job. Unlike traditional functions, a single Cloud Run container instance can process up to 1,000 concurrent requests simultaneously.

Instances execute inside secure gVisor sandbox environments and automatically scale from zero to hundreds of instances based on incoming request queue depth.

How pricing works once you are inside

Billed for allocated vCPU and RAM only while actively processing requests (CPU allocated during request processing), or on an always-allocated model for background services.

Google Cloud egress charges apply when sending data outside the Google network, which becomes a significant cost factor for high-bandwidth APIs.

Where it breaks down

Services requiring raw UDP traffic, privileged kernel capabilities, or persistent local block storage are outside the Cloud Run execution model.

What switching costs

Very low. Because applications are packaged as standard Docker containers listening on standard HTTP ports, migrating to AWS ECS, Kubernetes, or Fly is trivial.

Sources

Google Cloud Run architecture documentation and pricing calculators, October 2026.