Associate Cloud Engineer ยท Cloud Run operations and accelerators
Agent Runtime deployments
Deploying an agent to Agent Runtime on Gemini Enterprise Agent Platform (formerly Vertex AI Agent Engine): what it manages, and when to use it instead of Cloud Run.
Key points
- Agent Runtime, part of Gemini Enterprise Agent Platform (formerly Vertex AI Agent Engine), is a managed serverless runtime for deploying and scaling AI agents built with ADK, LangChain, LangGraph, AG2, LlamaIndex or custom code.
- It brings Sessions, Memory Bank, observability and enterprise security controls, so you do not build that plumbing on Cloud Run.
Source: https://docs.cloud.google.com/agent-builder/agent-engine/overview
Practice questions
Agent Runtime on Gemini Enterprise Agent Platform was formerly known as what?
- Vertex AI Agent Engine
- Cloud Functions Gen 1
- App Engine Flexible
- Dataflow Prime
Show answer
Answer: A. Vertex AI Agent Engine
It was previously Vertex AI Agent Engine. The other names are unrelated products.
Which capabilities does Agent Runtime offer to deployed agents?
- Sessions for conversation state
- GKE node pool autoscaling configuration
- Memory Bank for long-term memories
- Cloud SQL read replica management
- TPU slice scheduling
Show answer
Answer: A. Sessions for conversation state, C. Memory Bank for long-term memories
Sessions keep conversation state and Memory Bank stores long-term memories. GKE node pool tuning, Cloud SQL replicas and TPU slice scheduling are separate infrastructure concerns.
True or false: Agent Runtime can only run agents written with Google's Agent Development Kit.
Show answer
Answer: False
False. It also supports LangChain, LangGraph, AG2, LlamaIndex and custom agents.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
