# TODO(b/414798340): Modify this when PresetMetadata design is finalized. # TODO(b/436350694): Add boolean for presets with ingress container. // TODO(b/446719563): Use type and name instead of name and display_name. Use proper proto enums instead for type and data_type. presets: - name: ai-inference display_name: AI Inference version: 0.0.1 category: CATEGORY_QUICKSTART description: Create a service for running inference on AI models. supported_resources: - SERVICE config_values: CPU limit: 4 vCPUs Memory limit: 16GiB GPU: 1 NVIDIA L4 (no zonal redundancy) Billing: instance-based example_gcloud_usage: gcloud alpha run deploy --preset=ai-inference - name: ollama display_name: Ollama version: 0.0.1 category: CATEGORY_QUICKSTART description: Inference server for open LLMs, using GPUs and Cloud Storage. Deploys the latest Ollama container, configured for Cloud Run with L4 GPUs and a Cloud Storage bucket for model storage. supported_resources: - SERVICE parameters: - name: bucket label: GCS Bucket Name description: Connect your service to a Cloud Storage bucket. Models downloaded by Ollama will be stored in this bucket, improving overall performance and minimizing cold start times. type: GCS_BUCKET data_type: DATA_TYPE_STRING config_values: Container Image: ollama/ollama:latest CPU limit: 4 vCPUs Memory llimit: 16GiB GPU: 1 NVIDIA L4 (no zonal redundancy) Billing: instance-based example_gcloud_usage: - gcloud alpha run deploy --preset=ollama - gcloud alpha run deploy --preset=ollama:bucket= - name: private-service display_name: Private Service version: 0.0.1 category: CATEGORY_QUICKSTART description: Create a private, internal service with access control enforced based on identity and IAM roles. supported_resources: - SERVICE config_values: Ingress: internal Authentication: required example_gcloud_usage: gcloud alpha run deploy --preset=private-service - name: public-service display_name: Public Service version: 0.0.1 category: CATEGORY_QUICKSTART description: Publicly accessible endpoint, exposing your service to anyone on the internet, without any authentication. supported_resources: - SERVICE config_values: Ingress: all Authentication: allow public access example_gcloud_usage: gcloud alpha run deploy --preset=public-service - name: single-concurrency display_name: Single Concurrency version: 0.0.1 category: CATEGORY_QUICKSTART description: Create a service where requests are handled strictly one at a time. supported_resources: - SERVICE config_values: Container Concurrency: 1 example_gcloud_usage: gcloud alpha run deploy --preset=single-concurrency - name: static-website display_name: Static Website version: 0.0.1 category: CATEGORY_QUICKSTART description: Serve your static website assets from a Cloud Storage bucket with an NGINX server behind a Cloud Run endpoint, allowing you to use IAP, a load balancer or private networking. supported_resources: - SERVICE parameters: - name: bucket label: Storage Bucket Name description: Unique bucket name (and optional folder path) with static assets. type: GCS_BUCKET required: true data_type: DATA_TYPE_STRING config_values: Container Image: nginx:stable Storage Bucket: example_gcloud_usage: gcloud alpha run deploy --preset=static-website:bucket=