You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi there, I'm running into a bit of a blocker regarding agent configuration during deployment.
My goal is simple, I want to deploy an agent that terminates if it reaches a threshold of 5 LLM calls.
I initially solved this by using the max_llm_calls parameter inside RunConfig, which gets passed to the Runner. This works great for local interaction. However, as I move toward deployment (Vertex Agent Engine / Google Cloud Run), I'm required to use the App object.
I read here that today is not possible to use RunConfig with App for deployment, so what is the recommended pattern to enforce max_llm_calls in production?
Also is there a way to test a Runner specifically with adk web, or is the web interface strictly for Agent and App objects?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
Hi there, I'm running into a bit of a blocker regarding agent configuration during deployment.
My goal is simple, I want to deploy an agent that terminates if it reaches a threshold of 5 LLM calls.
I initially solved this by using the max_llm_calls parameter inside RunConfig, which gets passed to the Runner. This works great for local interaction. However, as I move toward deployment (Vertex Agent Engine / Google Cloud Run), I'm required to use the App object.
I read here that today is not possible to use RunConfig with App for deployment, so what is the recommended pattern to enforce max_llm_calls in production?
Also is there a way to test a Runner specifically with adk web, or is the web interface strictly for Agent and App objects?
Beta Was this translation helpful? Give feedback.
All reactions