Module assessment

1.

A software provider serves 40 small customers from a single Azure Cosmos DB container partitioned by a tenant identifier. None of the customers has a contractual or regulatory requirement for a separate account, but the operations team wants central reporting on usage per customer. What's the appropriate response?

Create a fleet and enroll the account, then use a throughput pool so each customer's traffic is billed separately.

Split each customer into a dedicated account, then group the accounts into a fleet so the estate can be managed centrally.

Keep the shared container and answer the reporting question from per-partition-key metrics and the application's own telemetry.

2.

A platform team enrolls an Azure Cosmos DB account into the production-eastus fleetspace of their retail fleet. A month later, they want that account to sit in a new production-eastus-premium fleetspace in the same fleet. What has to happen?

Add the account to the new fleetspace, which supersedes the earlier registration automatically.

Delete the account's fleetspace account registration in production-eastus, and then register it in the new fleetspace.

Delete and recreate the Azure Cosmos DB account, because fleetspace membership is fixed for the lifetime of the account.

3.

A container in a pooled fleetspace has one physical partition and an autoscale maximum of 8,000 RU/s. The fleetspace pool has a minimum of 200,000 RU/s and available capacity during a spike. What's the most that physical partition can consume?

10,000 RU/s, because a physical partition's combined dedicated and pooled consumption is capped.

13,000 RU/s, because a physical partition adds up to 5,000 RU/s from the pool to its dedicated throughput.

200,000 RU/s, because any resource in the fleetspace can consume the pool's full capacity.

4.

An operations team wants four accounts in one throughput pool: two single-region write accounts in East US, one single-region write account replicated to East US and West US, and one multi-region write account in East US. How many fleetspaces do these accounts require?

One, because all four accounts include East US in their regional configuration.

Three, because each distinct combination of regional configuration and service tier needs a fleetspace of its own.

Two, one for the single-region write accounts and one for the multi-region write account.

5.

To find the largest consumers of storage, an analyst queries FactResourceUsageHourly in a fleet analytics dataset. The results come back with correct storage figures, but every row is identified only by a ResourceId value. What resolves this problem?

Join the results to DimResource, which maps each ResourceId to a subscription, account name, region, database, and container.

Query FactAccountHourly instead, because resource usage figures are duplicated there with account names attached.

Re-enable the destination with account names included, because the exported columns are configurable per fleet.

Check your knowledge

Feedback