How can you use AI Gateway usage tables to monitor costs?

By setting up alerts in the AI Gateway dashboard

By viewing the model's training cost in the MLflow UI

By exporting the usage table to an external billing system

By querying the usage table to sum token usage and estimated costs over time

When enabling inference logging for a RAG application, what type of data is typically captured in the inference table?

Only error messages and failed requests

The model's internal weights and gradients during inference

The request payload, the response output, and metadata such as timestamps and model version

Only the final response text returned to the user

How do you implement rate limiting with AI Gateway?

By configuring rate limit policies on the AI Gateway endpoint

By adding a load balancer in front of the endpoint

By writing a custom script that rejects requests

By using the model serving UI to set limits

A company provides a free tier and a paid tier for their LLM API. Free-tier users should be limited to 10 requests per minute, while paid users should have a higher limit of 100 requests per minute. They use the AI Gateway. Which configuration should they use?

Set a single rate limit of 100 requests per minute for all users.

Configure two rate limiting policies, one for each tier, based on a user attribute that identifies the tier.

Set a single rate limit of 10 requests per minute for all users.

Use inference tables to track usage and manually block free users who exceed the limit.

A team has deployed a model through the AI Gateway. They want to set up an alert that triggers when the average cost per request exceeds a certain threshold over a 5-minute window. Which data source and approach should they use?

Use the inference table to calculate cost from the response payload.

Query the AI Gateway usage table and set up a scheduled job to compute the average cost and trigger an alert.

Use Agent Monitoring to track cost metrics.

Enable rate limiting to prevent cost overruns.

DatabricksCertified Generative AI Engineer Associate

Domain 6Objective 2

Monitoring and Logging for Deployed Applications GENERATIVE-AI-ENGINEER-ASSOCIATE Practice Questions (Page 1)

Part of the Section 6: Evaluation and Monitoring domain, which makes up ~17% of our current practice bank. Databricks does not publish an official question count, but from its 90-minute exam (~35–60 total, ~6–10 in this domain), expect 1–2 from this objective — we provide 25 practice questions to prepare you well beyond it. (estimate)

25questions here

5free pages

6concepts

Questions 1–5

1foundation · easy
How can you use AI Gateway usage tables to monitor costs?
Select an answer first
2foundation · easy
When enabling inference logging for a RAG application, what type of data is typically captured in the inference table?
Select an answer first
3foundation · easy
How do you implement rate limiting with AI Gateway?
Select an answer first
4expert · hard
A company provides a free tier and a paid tier for their LLM API. Free-tier users should be limited to 10 requests per minute, while paid users should have a higher limit of 100 requests per minute. They use the AI Gateway. Which configuration should they use?
Select an answer first
5application · medium
A team has deployed a model through the AI Gateway. They want to set up an alert that triggers when the average cost per request exceeds a certain threshold over a 5-minute window. Which data source and approach should they use?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to GENERATIVE-AI-ENGINEER-ASSOCIATE

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by Databricks. “GENERATIVE-AI-ENGINEER-ASSOCIATE” is a trademark of its owner, used for identification only.