Databricks-GenAI-Assoc Evaluation and Monitoring Practice Question
A team deploys a customer-support assistant on a Mosaic AI Model Serving endpoint and enables inference table logging to Unity Catalog. Compliance requires that every production response be traceable back to the exact request, the retrieved context, and the model version that produced it, and that reviewers can query this history with SQL months later. Which capability satisfies this requirement?
⚠ Common exam trap
Test-takers frequently confuse operational dashboards or client-side logs with auditable records, when only governed payload-level logging preserves request, context, and model version together.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
The endpoint's request and response payloads written to a Unity Catalog inference table, queried with SQL and joined to model version metadata.
Compliance-grade traceability demands durable, queryable records of each request, the context supplied to the model, the response, and the serving model version. Unity Catalog inference tables store exactly that as governed Delta tables, so SQL queries and joins to model version metadata answer audit questions months after the traffic occurred. Client logs, offline experiment runs, and aggregate dashboards all lack either the payload detail or the retention and governance needed.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
The endpoint's request and response payloads written to a Unity Catalog inference table, queried with SQL and joined to model version metadata.
Why this is correct
Inference tables persist the full request, retrieved context, response, and associated metadata as Delta tables governed by Unity Catalog, so reviewers can run SQL against historical traffic long after the fact. Because the logged rows include the served model version and timestamp, each response can be traced to its exact inputs and the model that generated it, which is precisely the auditability the compliance team requires.
- ✗
Real-time dashboards built on the endpoint's built-in metrics such as request count and error rate.
Why it's wrong here
Built-in endpoint metrics are aggregated operational counters; they contain no request text, no retrieved documents, and no per-response model version. Dashboards are designed for live health monitoring and typically retain only recent windows, which fails the requirement to query individual historical responses months later. They are complementary to audit logging, never a substitute for it.
- ✗
MLflow experiment runs recorded during offline evaluation of the assistant before release.
Why it's wrong here
MLflow experiments document evaluation runs on test datasets, not live production traffic. They contain metrics, parameters, and artifacts from a controlled experiment, so they cannot answer which model version served a specific customer request or what context was retrieved for it. Using them for compliance traceability would produce a record of pre-release testing rather than an auditable ledger of production responses.
- ✗
Client-side application logs streamed to a log analytics workspace and retained for thirty days.
Why it's wrong here
Application logs typically capture errors and timing rather than the complete retrieved context and model version, and a thirty-day retention window cannot satisfy a requirement to query history months later. They also live outside Unity Catalog governance, so access control and SQL-based auditing over the same tables as the data are not available. This approach is useful for debugging but insufficient for compliance-grade traceability.
About these practice questions
Courseiva writes every Databricks-GenAI-Assoc question from scratch — 330 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Databricks exam blueprint
This Databricks-GenAI-Assoc practice question is part of Courseiva's free Databricks certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the Databricks-GenAI-Assoc exam.