What is the default timeout when a predict request is made to an LLM model deployed on vertex AI

[Yash2384]

I am using this library to make a prediction request to the model deployed on Vertex AI. I am getting a timeout exception, Not sure if I need to increase the timeout and up to what value . Also what is the default value , I can find nothing in the documentation 

Client API  : https://cloud.google.com/python/docs/reference/aiplatform/latest/google.cloud.aiplatform_v1.services...

[maricarp · Accepted solution]

Hello,

As quoted in the documentation, 

---

---

“Requests timeout after 60 seconds for both public and private endpoints.”

---

If you would like to request a timeout more than 60 seconds, you must file a support ticket or contact your Google Cloud representative.

I hope this helps.

[davixie]

In my case, I was deploying the endpoint as a shared public endpoint, and according to the documentation , shared public endpoints have a timeout of 60 seconds. So, if you need a timeout longer than 60 seconds, you need to deploy it as a dedicated endpoint.