Which Databricks feature records requests to a live endpoint?
Why D is the answer
Inference tables
Inference tables record the requests and responses of a live endpoint, so you can query and evaluate them.
In detail
- Inference tables record requests and responses for a live endpoint.
- You can then query and evaluate them.
- Databricks now documents these as Unity Gateway inference tables.
- Treat these tables as sensitive, because they hold real user text.
Use case
An illustrative scenario, not a real incident.
A team wants to review what users ask a live endpoint. Inference tables record requests and responses, and analysts query them like any table.
See it in different ways
flowchart LR A["Requests"] --> B["AI Gateway"] B --> C["Rate limits"] B --> D["Inference tables"] B --> E["Usage tables"] D --> F["Agent Monitoring"]
sequenceDiagram participant U as Caller participant G as AI Gateway participant E as Endpoint participant T as Inference tables U->>G: Request G->>G: Apply rate limits G->>E: Forward E-->>U: Response E->>T: Record the request and response
flowchart LR S0["Request"] --> S1["Endpoint"] --> S2["Response"] --> S3["Inference table"]
Key point: Inference tables record live requests.
Why the other options are wrong
- A. Cluster logs only
- Cluster logs do not give you requests as data.
- B. System tables for billing
- Billing tables show usage cost, not payloads.
- C. Delta table history
- Table history records changes to a table.