Google Cloud Run

What Google Cloud Run is

Google Cloud Run is a managed application platform for running code, functions and containers on Google's infrastructure. Workloads can run as services, jobs, worker pools or instances, and containers scale up or down automatically, including to zero. For supported languages, Cloud Run can build and deploy source code with Google Cloud buildpacks without a Dockerfile. It also supports hosting large language models and real-time AI inference using GPUs. The platform is listed for API, Linux and web, with recommended integrations spanning build and deployment, observability, databases, storage and other Google Cloud services. Cloud Run has a free plan and free trial; new customers receive $300 in credits for Cloud Run and other Google Cloud products. Pay-as-you-go billing charges CPU at $0.00001800 per vCPU-second and memory at $0.00000200 per GiB-second beyond the monthly free tier. The free tier includes the first 240,000 vCPU-seconds and 450,000 GiB-seconds per month; pricing varies by region and billing configuration. Requests can last up to 60 minutes, while each container instance can use up to 32 GiB of memory and 8 vCPU. Containers must use the Linux x86_64 ABI. Google says service endpoints use TLS termination and encrypted channels to the appropriate region, and container isolation uses Linux namespaces and security profiles.

Who it suits

It suits developers and teams running code, functions or containers on Google Cloud, including supported AI workloads. Its free tier, autoscaling and service, job, worker-pool and instance options may fit varied workloads.

Limits

Each request can run for up to 60 minutes, and a container instance can use up to 32 GiB of memory and 8 vCPUs. The free tier includes 240,000 vCPU-seconds and 450,000 GiB-seconds per month; usage beyond those allowances is billed.

Getting started

1. Open the Cloud Run website
2. Choose a service, job, worker pool, or instance workload
3. Deploy a supported source language with buildpacks or provide a Linux x86_64 container
4. Connect required Google Cloud integrations
5. Use the free plan or pay-as-you-go billing
6. Monitor the deployed workload through Google Cloud

Verdict

Cloud Run offers several workload forms, automatic scaling and source deployment options, with a free plan and trial credits. Check the ABI, request-duration and instance limits, and account for regional variation in pay-as-you-go charges.