What should you do?

You have a container named container1 in an Azure Cosmos DB for NoSQL account named account1 that is set to the session default consistency level. The average size of an item in container1 is 20 KB.

You have an application named App1 that uses the Azure Cosmos DB SDK and performs a point read on the same set of items in container1 every minute.

You need to minimize the consumption of the request units (RUs) associated to the reads by App1.

What should you do?
A . In account1, change the default consistency level to bounded staleness.
B. In App1, change the consistency level of read requests to consistent prefix.
C. In account1, provision a dedicated gateway and integrated cache
D. In App1, modify the connection policy settings.

View Answer

Answer: B

Explanation:

The cost of a point read for a 1 KB item is 1 RU. The cost of other operations depends on factors such as item size, indexing policy, consistency level, and query complexity1. To minimize the consumption of RUs, you can optimize these factors according to your application needs.

For your scenario, one possible way to minimize the consumption of RUs associated to the reads by App1 is to change the consistency level of read requests to consistent prefix. Consistent prefix is a lower consistency level than session, which is the default consistency level for Azure Cosmos DB. Lower consistency levels consume fewer RUs than higher consistency levels2. Consistent prefix guarantees that reads never see out-of-order writes and that monotonic reads are preserved1. This may be suitable for your application if you can tolerate some eventual consistency.