DP-900 Practice Question: Describe considerations for working with non-relational data on Azure

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 default consistency level is Strong.

  Why it's wrong here

  Azure Cosmos DB offers multiple consistency levels, including Strong, Bounded Staleness, Session, Consistent Prefix, and Eventual, but the default is Session, not Strong. Session consistency guarantees monotonic reads and writes for a single client session, providing a practical balance between performance and data consistency. Strong consistency is available but comes with higher read latency and is not supported with multi-region write configurations, making it unsuitable as the default for most workloads.

• ✗

  It uses DTUs to measure performance.

  Why it's wrong here

  Azure Cosmos DB measures performance using Request Units (RUs) per second, not DTUs. DTUs (Database Transaction Units) are a partially abstracted metric used by Azure SQL Database to represent a blended mix of CPU, memory, and I/O capacity. In Cosmos DB, each operation's resource consumption depends on factors like item size, indexing, consistency level, and query complexity, so RU provisioning accurately reflects the cost of each request and enables fine-grained scaling.

• ✓

  It guarantees single-digit millisecond latency for reads and writes at the 99th percentile.

  Why this is correct

  Azure Cosmos DB's globally distributed, multi-master architecture and automatic indexing capabilities are fundamental to its ability to guarantee single-digit millisecond latency for reads and writes at the 99th percentile. This performance commitment is enshrined in its financially-backed Service Level Agreement (SLA), ensuring predictable and consistent high-speed data access. This accurately describes a core truth about Azure Cosmos DB, directly satisfying the question's requirement to identify a true characteristic.

• ✗

  It is a relational database management system.

  Why it's wrong here

  Cosmos DB is a NoSQL database that uses a schema-agnostic, multi-model approach with APIs for document, graph, and key-value data, whereas a relational database management system enforces a fixed schema with ACID transactions and SQL-based joins. It is tempting because Cosmos DB offers a SQL API that mimics relational query syntax, but it still lacks relational integrity constraints and normalised table structures, making it the correct choice only for globally distributed, low-latency workloads that do not require strict relational modelling.

• ✓

  It supports multiple data models including document, key-value, graph, and column-family.

  Why this is correct

  Azure Cosmos DB is a multi-model NoSQL database that natively supports document, key-value, graph, and column-family data models through its diverse APIs. You can interact with the same underlying storage using the SQL API for documents, the Cassandra API for wide-column data, the Gremlin API for graphs, and the Table API for key-value stores. This breadth allows you to employ a single globally distributed service across heterogeneous application patterns without needing separate database engines.

Go deeper

Related to this question