# Practical and Useful Patterns with ADK Deploy an agent without writing a single Dockerfile or K8s manifest? ☁️ With `adk deploy cloud_run`, your agent goes live on container serverless in one command. Add Dev UI too for seamless development and operations! ## 📌 Title Cloud Run — One-Command Container Serverless Deployment ## 🔗 URL ## 🧩 Overview ADK's Cloud Run deployment uses the `adk deploy cloud_run` command to deploy agents to a serverless environment without manual Dockerfile or manifest creation. Use `--with_ui` to bundle the Dev UI, `--session_service_uri=agentengine://` to integrate with managed sessions, and get an auto-scaling web backend ready to go. ## 🛠 How to Use Basic deployment command: ```bash adk deploy cloud_run \ --project=$GOOGLE_CLOUD_PROJECT \ --region=$GOOGLE_CLOUD_LOCATION \ $AGENT_PATH ``` Key options: - `--with_ui`: Deploy the ADK development interface alongside the API server - `--session_service_uri`: URI of the session service (pass `agentengine://` for Agent Runtime's managed sessions) - `--service_name`: Customize the Cloud Run service name - `--port`: Specify container listening port (default: 8000) Alternatively, use `gcloud run deploy` with custom Dockerfiles to embed agents in existing FastAPI applications. ## 🏗 Practical Usage - **Sharing Prototypes**: Deploy with `--with_ui` to let your team try the agent via the Dev UI - **Existing Web App Integration**: Use `gcloud run deploy` to add agent capabilities to custom FastAPI apps - **Managed Sessions**: Use `--session_service_uri=agentengine://` for Agent Runtime session management, enabling stateful conversations - **Auto-scaling**: Traffic-based auto-scaling for stable web backend operations - **Auth Control**: Choose whether to allow unauthenticated access at deploy time. For private APIs, access via `gcloud auth print-identity-token` ## 💡 Use Cases - Quickly deploying agents with Dev UI as internal tools - Adding agent features to existing FastAPI/Flask applications - Running Slack bots or chat app backends in serverless mode - Building stateful chat services with Agent Runtime managed sessions ## ⚠️ Caveats - You'll be prompted about unauthenticated access during deployment — only choose 'y' for public APIs - Consider setting minimum instances for latency-sensitive workloads to mitigate cold starts - `--with_ui` is for development — configure proper authentication for production - Default port is 8000 — use `--port` to change if it conflicts with existing services ## ✨ No Dockerfile, no manifest. One command to ship a serverless agent to the world. Experience the simplicity of Cloud Run deployment! ⚡ #ADK# #AIAgent#

OpenShell v0.0.40 🔀 local-domain service routing in the gateway ☸️ k8s node scheduling + tolerations 🔒 CLI TLS now uses the OS trust store 🛡️ SecretResolver debug no longer leaks secrets Two security fixes ship alongside new routing and K8s scheduling control.