Databricks-GenAI-Assoc Assembling and Deploying Apps Practice Question
A platform team is preparing to deploy a GenAI chat application built on Mosaic AI Model Serving. They want the deployment to enforce least-privilege access for the application and to keep the serving environment reproducible. Which two practices should they implement? (Choose two.)
⚠ Common exam trap
The trap here is equating 'convenient deployment' with broad admin rights or always-latest dependencies, when least privilege and pinned versions are what the scenario actually demands.
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
✓
Pin the model's dependency versions in the logged MLflow environment so the serving environment installs a known, reproducible set of packages.
Least privilege is achieved by granting the application principal only endpoint query rights plus the minimum Unity Catalog reads for the model and its data. Reproducibility is achieved by pinning dependency versions in the logged MLflow environment so the serving container installs a known package set. Workspace admin, dropping Unity Catalog, and always-latest installs each trade away security or reproducibility, so they do not meet the two stated goals.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Allow the endpoint to install the latest version of every dependency at startup so the environment always has current patches.
Why it's wrong here
Installing latest versions at startup makes deployments non-reproducible because the environment can change between restarts or replica scale-ups. A new upstream release can introduce breaking changes that take down the agent without any code change. This conflicts with the reproducibility requirement and should be replaced by pinned versions with deliberate upgrades.
- ✗
Give the application principal workspace admin so it can create endpoints and modify cluster policies as needed during deployment.
Why it's wrong here
Granting workspace admin violates least privilege and gives the application far more power than querying an endpoint requires. A compromised app with admin rights could alter workspace settings, policies, and other users' resources. Deployment-time endpoint creation should be performed by a deployment identity, not by the runtime application principal.
- ✓
Pin the model's dependency versions in the logged MLflow environment so the serving environment installs a known, reproducible set of packages.
Why this is correct
Pinning dependencies in the model's logged environment ensures the serving container installs the same package versions used during development, which makes the deployment reproducible and avoids version drift that can break the agent. This directly supports the reproducibility objective and prevents surprises when the endpoint restarts or scales.
- ✓
Grant the application's service principal only the endpoint query permission and the minimum Unity Catalog privileges needed to read the model and its dependencies.
Why this is correct
Least privilege means the application principal should have exactly the permissions required: CAN QUERY on the serving endpoint and read access to the model and referenced data assets in Unity Catalog. This limits blast radius if credentials leak and satisfies the security goal without granting broad workspace admin rights that the app does not need.
- ✗
Disable Unity Catalog for the model and store it in the workspace model registry to simplify permission management.
Why it's wrong here
Moving away from Unity Catalog removes centralized governance, lineage, and fine-grained access control, which undermines security rather than improving it. Workspace registry lacks the same privilege model and cross-workspace consistency. This contradicts the least-privilege goal and makes auditing model access harder.
About these practice questions
This Databricks-GenAI-Assoc question is part of Courseiva's 330-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam 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.