Workspace roles, XMLA endpoint and monitoring in Fabric

DP-600 · Maintain a data analytics solution · Updated

Fabric exam questions often ask who may do what and how to see why something is slow. Learn the role model, the XMLA setting and the two monitoring tools.

Workspace roles

RoleCan do
AdminEverything, including managing access and deleting the workspace
MemberEdit, publish and share items, add people (except Admins)
ContributorCreate and edit items, write data; cannot manage access
ViewerView and read items; data access in a warehouse still follows item permissions

Use security groups for roles, and keep one Admin if the scenario says so. A workspace identity is a service principal for the workspace, used to authenticate to resources such as firewall-protected storage. It is not a user role.

Item-level and data security

Roles apply to the whole workspace. For narrower access, share the item and then secure the data:

  • Warehouse: T-SQL GRANT, DENY, column-level security, row-level security, and dynamic data masking.
  • Semantic model: row-level security roles and object-level security, plus the Build permission for reuse.
  • Lakehouse: OneLake data access roles for folder-level control; the SQL analytics endpoint is read only.
GRANT SELECT ON OBJECT::dbo.Sales TO [user1@contoso.com];

XMLA endpoint

Tools like Tabular Editor, SSMS and ALM Toolkit connect to a semantic model through its XMLA endpoint. In the capacity settings, set XMLA Endpoint to Read Write to allow publishing. Read only allows queries.

For unattended access, use a service principal that is allowed in the tenant settings and added to the workspace. Large models deploy and refresh better with the Large semantic model storage format.

Monitoring

ToolUse it for
Monitoring hubStatus of pipeline, notebook, dataflow and refresh runs
Capacity Metrics appCapacity units used, smoothing, throttling and overload
DMVs (sys.dm_exec_requests)Live query detail in a warehouse
Query insights viewsLong-running and frequent queries in a warehouse

Throttling happens when consumption exceeds the capacity after smoothing. Only the Capacity Metrics app answers “is the capacity throttling”.

Exam traps

  • A workspace role is broad. Fine-grained control on warehouse tables is done with T-SQL permissions.
  • Contributor cannot manage access. Member can add users, Contributor cannot.
  • Read only XMLA is the default. Publishing needs Read Write.
  • Monitoring hub shows runs, not capacity throttling.
  • A workspace identity is for outbound authentication, not for giving users access.

Practice questions

Choose an answer, then review the explanation to check your understanding.

1. A user must read one table in a warehouse and nothing else in the workspace. What is the best first step?

Show answer

Answer: B.Sharing the item gives the connection, then T-SQL GRANT on the table limits access. Workspace roles are too broad.

2. Tabular Editor must publish changes directly to a semantic model in a Fabric workspace. What must be enabled?

Show answer

Answer: C.Writing through the XMLA endpoint needs the capacity setting XMLA Endpoint = Read Write. The default is read only.

3. Warehouse queries slow down several times a day and you suspect capacity throttling. What do you use to confirm it?

Show answer

Answer: B.The Capacity Metrics app shows utilization, smoothing and throttling per capacity. The Monitoring hub shows individual activity runs.

4. Developers must create and edit items but must not manage access, and you want minimal administrative effort. What do you do?

Show answer

Answer: B.Assign roles to a group, not to individuals. Contributor can create and edit items but cannot manage access.

DP-600 course outline

All 6 lessons · Course overview
  1. Git integration and deployment pipelines in FabricMaintain a data analytics solution Free
  2. Workspace roles, XMLA endpoint and monitoring in FabricMaintain a data analytics solution Free
  3. Lakehouse or warehouse and loading data in FabricPrepare data Full access
  4. Delta tables, Spark and Power Query transformationsPrepare data Full access
  5. Direct Lake and storage modes for semantic modelsImplement and manage semantic models Full access
  6. Model design, DAX and dynamic formattingImplement and manage semantic models Full access