RunR vs Google Cloud Run
RunR is an open-source (MIT) serverless platform: functions, multi-file projects and containers on Knative, on Kubernetes clusters you operate. Google Cloud Run is a fully managed Google Cloud platform that runs code, functions or containers as services, jobs, worker pools and (in Preview) instances.
Both run request-driven containers that scale to zero and shift traffic between revisions. The main difference is where they run and who operates them: Cloud Run runs in Google Cloud regions and Google operates it; RunR runs on your clusters and you operate it.
Differences at a glance#
| RunR | Google Cloud Run | |
|---|---|---|
| Where it runs | Your Kubernetes clusters (1.30 or newer), each connected as a region; bare metal, a private cloud or managed Kubernetes (Regions) | Google Cloud regions; each Cloud Run resource resides in one region (locations) |
| Who operates it | You: the control plane installs with Helm or Docker Compose (Production install, Single-host install) | Google: a fully managed platform (what is Cloud Run) |
| From source | Web editor, git push or the CLI; images are built in the cluster with Kaniko (Git deploy, CLI) |
gcloud run deploy --source builds with Google Cloud's buildpacks and Cloud Build, or with your Dockerfile (deploying from source) |
| Scale to zero | On by default for every project, with minimum and maximum scale and a timed warmer (Scaling) | Services scale to zero by default; minimum instances default to 0 (autoscaling, minimum instances) |
| Rollouts | Blue-green with a test URL; canary in steps that rolls back by itself on error rate or p95 latency (Deploy strategies) | Traffic split by percentage between revisions, gradual rollout and rollback (rollouts and traffic migration) |
| Isolation | runc, gVisor or Kata Containers per project type, set up per region from the console (Kata Containers hypervisors) | First generation based on gVisor; second generation a microVM (execution environments) |
| Events | HTTP, schedules, queues, webhooks, topics, Kafka, NATS, RabbitMQ, MQTT, Postgres CDC, S3 and IMAP, with retries and dead letters (Triggers) | Eventarc events, Pub/Sub push subscriptions and Cloud Scheduler (Eventarc, Pub/Sub, Cloud Scheduler) |
| Several regions | Global projects: one host name, HA or proximity mode, failover between regions (Global projects) | A global external Application Load Balancer with serverless NEGs, or multi-region services with automated failover (multiple regions) |
| GPUs | GPU functions and containers on the GPU nodes of your clusters; Hugging Face model services (GPU workers, Hugging Face models) | NVIDIA L4 and RTX PRO 6000 Blackwell, one GPU per instance, in selected regions (GPU) |
| Workflows | Durable workflows built in, with Python and TypeScript SDKs (Durable workflows) | Google Cloud Workflows, a separate orchestration service that calls Cloud Run over HTTP (Workflows) |
| Trusted images | SBOMs, scans, cosign signatures, a policy gate and signed-image admission (Supply chain) | Binary Authorization deploys only trusted container images (Binary Authorization) |
| Infrastructure as code | Terraform / OpenTofu provider brf-tech/runr, the runr CLI, a GitHub Action (Terraform provider) |
Terraform resource google_cloud_run_v2_service of the google provider (registry) |
Cloud Run's v1 Admin API Service resource is compatible with the Knative Serving API (Cloud Run APIs); RunR deploys its projects as Knative Services on Knative Serving (Architecture).
When to choose which#
Google Cloud Run fits when your workloads already live on Google Cloud, you want Google to run the platform, and you want its regions, GPUs and Google Cloud event sources without operating clusters.
RunR fits when code, data and logs must stay on infrastructure you operate (on premises, a private cloud, air-gapped regions), when you already run Kubernetes and want one platform across your clusters, or when many teams share it with organizations, roles and single sign-on (Roles and permissions, Modules).
RunR documentation to read next#
- Production install and single-host install
- Regions: what a cluster needs and how to connect it
- Triggers and Scaling
- Deploy strategies and Global projects
- All comparisons
The statements about Google Cloud Run come from Google's documentation, read on 2026-11-03 and linked next to each one. Google Cloud and Cloud Run are trademarks of Google LLC. RunR is not affiliated with or endorsed by Google.