Medium — DP-420 Practice Quiz

Question 1

A retail company runs an Azure Cosmos DB account with a single write region in East US. Their customer base has expanded significantly to Western Europe, and European users are reporting slow write latency when submitting orders, while reads remain acceptable. The team wants to reduce write latency for European customers with minimal application changes and without sacrificing the ability to resolve conflicting concurrent writes. Which configuration change should they make?

Accepted Answer

Enable multi-region writes (multi-region writes) and add West Europe as an additional write region.

Answer

Add West Europe as a read region only and enable automatic failover with West Europe as the highest failover priority.

Answer

Change the consistency level from session to eventual to reduce write acknowledgment time.

Answer

Perform a manual failover to make West Europe the single write region instead of East US.

Question 2

A retail company exposes its Azure Cosmos DB for NoSQL product catalog through an Azure AI Search index using an indexer. Products are frequently discontinued, and when they are, the corresponding document is physically deleted from the Cosmos DB container. Users complain that discontinued products still appear in search results even after the indexer runs on schedule. Which action should you take so the indexer removes deleted products from the search index?

Accepted Answer

Enable the indexer's soft-delete deletion detection policy and add a soft-delete marker field to documents instead of physically deleting them.

Answer

Enable the high water mark change detection policy using the _ts field on the Cosmos DB documents.

Answer

Reset and re-run the indexer on a more frequent schedule so it performs a full re-crawl of the container each time.

Answer

Configure a field mapping that maps the Cosmos DB id property to the search index key field.

Question 3

A retail company stores product documents in an Azure Cosmos DB for NoSQL container. The product catalog team wants to add a semantic product search feature to their website that returns products based on meaning rather than exact keyword matches. They plan to use Azure AI Search over the Cosmos DB data. The team already generates text embeddings for each product description. Which Azure AI Search capability should they configure to satisfy this semantic-similarity requirement?

Accepted Answer

Define a vector field in the index and use the Cosmos DB indexer with a field mapping to store the precomputed embeddings, then issue vector queries against that field.

Answer

Enable the built-in Cosmos DB analytical store and query it directly from Azure AI Search using T-SQL.

Answer

Configure a scoring profile that boosts documents based on the freshness of the product's last-modified timestamp.

Answer

Create a synonym map that lists related product terms so the search engine can expand keyword queries automatically.

Question 4

A healthcare application stores patient records in Azure Cosmos DB for NoSQL. Compliance requires that the 'nationalId' property be encrypted client-side so that plaintext is never visible to the database engine. Developers must still be able to run queries that filter documents using an exact-match equality predicate on 'nationalId'. When configuring Always Encrypted for this property, which encryption type must you select?

Accepted Answer

Deterministic encryption

Answer

Randomized encryption

Answer

Service-managed key encryption at rest

Answer

Transparent Data Encryption (TDE) with a customer-managed key

Question 5

You maintain a Cosmos DB container of blog posts where each document has a 'tags' property that is an array of strings, for example ["azure", "cosmosdb", "nosql"]. Your application must return all posts that include the tag "azure" anywhere in the array. You want a single SQL query expression that correctly filters these documents. Which WHERE clause should you use?

Accepted Answer

WHERE ARRAY_CONTAINS(c.tags, "azure")

Answer

WHERE c.tags = "azure"

Answer

WHERE ARRAY_LENGTH(c.tags) > 0 AND c.tags[0] = "azure"

Answer

WHERE CONTAINS(c.tags, "azure")

Question 6

A financial services company runs a single-write-region Cosmos DB account with the write region in East US and read regions in West US and North Europe. They have enabled service-managed (automatic) failover and configured failover priorities as: 0 = East US, 1 = North Europe, 2 = West US. During a regional outage that makes East US unavailable, what happens to the write region?

Accepted Answer

Cosmos DB automatically promotes North Europe to the new write region based on the configured failover priority.

Answer

Writes are unavailable until an administrator manually triggers a failover to a healthy region.

Answer

Cosmos DB automatically promotes West US to the new write region because it has the lowest network latency to clients.

Answer

The account is converted to a multi-region write account so all remaining regions accept writes.

Question 7

You manage a Cosmos DB container for an e-commerce catalog. Traffic is highly variable: it sits near 800 RU/s for most of the day but spikes to roughly 9,000 RU/s for two to three unpredictable hours during flash sales. The team wants to avoid throttling during spikes while minimizing cost during idle periods, and they do not want to manually adjust throughput or write scaling code. Which throughput configuration should you choose?

Accepted Answer

Configure autoscale throughput with a maximum of 10,000 RU/s so the container scales between 1,000 and 10,000 RU/s automatically.

Answer

Provision standard (manual) throughput at 9,000 RU/s to guarantee capacity for the spikes.

Answer

Use serverless mode so you only pay per request consumed with no capacity planning.

Answer

Provision standard (manual) throughput at 1,000 RU/s and rely on burst capacity to absorb the flash-sale spikes.

Question 8

A retail team stores a product catalog in an Azure Cosmos DB for NoSQL container. They want to expose full-text and faceted search over the catalog using Azure AI Search, and the index must automatically pick up new and modified products without re-crawling the entire container each run. Products are updated frequently throughout the day. What is the most appropriate way to configure the Azure AI Search indexer for this requirement?

Accepted Answer

Configure an Azure AI Search indexer with a change detection policy based on the container's _ts property so the indexer processes only documents changed since the last run.

Answer

Schedule the indexer to run a full reset and re-index of all documents on every execution to guarantee the index stays current.

Answer

Push documents into Azure AI Search directly from the application on every write instead of using an indexer.

Answer

Enable the analytical store on the container and point the Azure AI Search indexer at the analytical store for change tracking.

Question 9

An e-commerce platform stores individual order documents in a Cosmos DB container partitioned by /customerId. The analytics team needs a separate 'customerSummary' container that maintains a running total of each customer's lifetime spend and order count, updated within seconds of any new order. The solution must scale automatically and avoid expensive periodic full-container scans. Which approach best meets these requirements?

Accepted Answer

Deploy an Azure Function with a Cosmos DB trigger bound to the orders container's change feed that reads each changed order and upserts the aggregated totals into the customerSummary container.

Answer

Schedule a timer-triggered Azure Function that runs a cross-partition GROUP BY query over the entire orders container every minute and rewrites the customerSummary container.

Answer

Enable analytical store on the orders container and query it with Azure Synapse Link to compute the aggregates on demand for the summary.

Answer

Create a post-trigger on the orders container that recomputes each customer's totals by querying all their orders whenever a new order is inserted.

Question 10

An e-commerce platform stores each customer's order in an 'orders' container partitioned by /customerId. The team needs an Azure Function triggered by the Cosmos DB change feed that, for every new or updated order, must (1) write a denormalized summary document to a 'customerOrderHistory' container and (2) push a message to a Service Bus queue for downstream fulfillment. The function must process changes reliably and scale across partitions automatically. Which approach correctly implements this integration?

Accepted Answer

Use the Cosmos DB trigger with a lease container, and within the same function bind an output to the 'customerOrderHistory' container and a Service Bus output binding, so each change is fanned out to both destinations.

Answer

Use a Timer trigger that queries the orders container every minute for recently modified documents, then writes to both destinations, since the change feed cannot invoke multiple outputs.

Answer

Use the Cosmos DB trigger but disable the lease container to avoid duplicate processing, and write only to 'customerOrderHistory' because a single function cannot have more than one output binding.

Answer

Use an HTTP trigger invoked by a change-feed pull-model loop in a separate service, then use the SDK directly to write to both destinations because output bindings are not supported for Cosmos DB writes.

Question 11

Your team built an Azure Function with a Cosmos DB trigger to process the change feed of an 'Orders' container for real-time fraud scoring. During a marketing campaign, order volume increases tenfold. You observe that a single Function instance is processing all changes and cannot keep up, causing rising latency. The container has 20 physical partitions. What is the correct way to enable the Function to scale out and process changes in parallel across multiple instances?

Accepted Answer

Ensure a lease container is configured for the trigger; the Consumption plan will automatically scale out Function instances, distributing partition key ranges (leases) across them.

Answer

Increase the MaxItemsPerInvocation setting so a single instance reads larger batches, which removes the need for additional instances.

Answer

Switch the Cosmos DB account to multi-region writes so each region's Function app processes a subset of the change feed independently.

Answer

Manually create 20 separate Function apps, one per physical partition, and hard-code each to poll a specific partition key range.

Question 12

You operate a production Azure Cosmos DB for NoSQL account with a single container using autoscale throughput. Users intermittently report HTTP 429 errors during peak hours. You want to be proactively notified the moment the underlying cause—a physical partition consistently approaching its throughput ceiling—begins to occur, before users experience widespread throttling. Which Azure Monitor metric should you build the alert rule on to detect this specific condition?

Accepted Answer

Normalized RU Consumption, alerting when it approaches 100%

Answer

Total Request Units, alerting when the sum exceeds the provisioned maximum

Answer

Server Side Latency, alerting when average latency exceeds 10 ms

Answer

Total Requests, alerting when the count of 200 status codes drops

Question 13

An e-commerce team needs to ingest 500,000 order documents nightly into a Cosmos DB container partitioned by /customerId. The documents span thousands of different customerId values, and each document is independent — partial failures are acceptable and should simply be retried. The team wants maximum ingestion throughput with the least code complexity. Which approach should they use?

Accepted Answer

Enable bulk execution (AllowBulkExecution = true) in the SDK and issue concurrent CreateItemAsync calls

Answer

Group the documents and use TransactionalBatch operations across the different customerId values

Answer

Wrap all inserts in a single stored procedure executed once per night

Answer

Insert documents sequentially with individual CreateItemAsync calls to preserve ordering

Question 14

A financial services company runs a globally distributed Cosmos DB account with a single write region in East US and read regions in West Europe and Southeast Asia. Compliance requires that reads never lag the write region by more than a defined number of write operations OR a defined time interval, whichever comes first, so auditors can bound the maximum data divergence. Which consistency level satisfies this requirement, and what must the team configure?

Accepted Answer

Bounded staleness, configuring the maximum lag as both a number of versions (K) and a time interval

Answer

Session consistency, configuring a session token TTL to bound divergence

Answer

Consistent prefix, configuring a maximum prefix length per region

Answer

Strong consistency, configuring the acceptable replication window in milliseconds

Question 15

A developer is writing a .NET application to migrate 50 million JSON documents into an Azure Cosmos DB for NoSQL container that has been provisioned with autoscale up to 40,000 RU/s. The migration runs from an Azure VM in the same region as the Cosmos DB account and must complete as quickly as possible while maximizing throughput utilization. Which combination of SDK configuration settings should the developer use?

Accepted Answer

Enable Bulk mode by setting AllowBulkExecution = true on CosmosClientOptions, and use Direct connection mode.

Answer

Use Gateway connection mode with AllowBulkExecution = true to reduce the number of open TCP connections.

Answer

Wrap all inserts in a single stored procedure and call it repeatedly to guarantee transactional consistency across the import.

Answer

Issue individual CreateItemAsync calls sequentially with Direct connection mode and rely on automatic retries for throttled requests.

Question 16

A retail company stores order documents in an Azure Cosmos DB for NoSQL container. Compliance requires that once an order document reaches a 'Completed' status, a copy must be moved to Azure Blob Storage for long-term archival, and the archived record must never be lost even if processing fails temporarily. The team wants a serverless, low-maintenance approach that automatically reacts to changes. Which design best meets these requirements?

Accepted Answer

Deploy an Azure Function with a Cosmos DB trigger (change feed) that writes completed orders to Blob Storage, and rely on the lease container and built-in checkpointing/retry to guarantee each change is processed.

Answer

Schedule an Azure Function on a timer trigger every 5 minutes to query the container for completed orders and write them to Blob Storage.

Answer

Enable analytical store with Azure Synapse Link and run a nightly Spark job that exports completed orders to Blob Storage.

Answer

Configure a post-trigger on the container that writes each completed order directly to Blob Storage during the write operation.

Question 17

Your team runs an Azure Functions app with a Cosmos DB trigger that processes the change feed from an 'Orders' container. During a recent sales event, downstream systems reported that order-processing events were arriving several minutes late. You need to programmatically measure how far behind the change feed processor is falling so you can trigger alerts and scale the function before the backlog grows. Which approach should you implement?

Accepted Answer

Use the change feed estimator to report the number of remaining (unprocessed) items per lease, and emit that value as a custom metric.

Answer

Query the lease container directly and count the number of lease documents to determine processing lag.

Answer

Enable the integrated cache on a dedicated gateway so change feed reads are served faster and lag is eliminated.

Answer

Increase the container's provisioned throughput (RU/s) until the reported 429 rate reaches zero.

Question 18

An e-commerce platform stores Customers and Orders in separate Azure Cosmos DB for NoSQL containers. When a customer record is soft-deleted (a 'deleted' flag set to true), the business requires that all of that customer's orders be automatically marked as orphaned within seconds, without impacting the write latency of the customer-facing application. Which approach best satisfies this requirement?

Accepted Answer

Deploy an Azure Function with a Cosmos DB trigger on the Customers container that reacts to change feed events and updates the related orders in the Orders container.

Answer

Add a post-trigger to the Customers container that queries and updates all matching order documents in the Orders container within the same operation.

Answer

Enable Azure Synapse Link on the Customers container and run a nightly Spark job to reconcile orphaned orders in the analytical store.

Answer

Configure a stored procedure on the Customers container that transactionally updates both the customer and all their orders in a single batch.

Question 19

An e-commerce team stores orders in a Cosmos DB container partitioned by /customerId. A separate reporting service needs a continuously updated materialized view aggregating daily sales totals per product into a second container partitioned by /productId. The team wants a low-latency, resilient mechanism that reacts to each new or updated order without polling or scheduled full scans. Which approach should they implement?

Accepted Answer

Use the Azure Cosmos DB change feed (via the change feed processor or an Azure Function trigger) to read order changes and write aggregates to the product container.

Answer

Create a post-trigger on the orders container that writes to the product container within the same operation.

Answer

Schedule a nightly cross-partition query with the SDK to scan all orders and rebuild the product aggregates.

Answer

Enable a stored procedure that runs on a timer to aggregate orders and upsert into the product container.

Question 20

A logistics company stores shipment documents in a Cosmos DB for NoSQL container. A dashboard runs a frequent query that filters on `WHERE c.warehouseId = @id` and returns results ordered by two fields: `ORDER BY c.status ASC, c.createdDate DESC`. The query currently fails at runtime with an error stating that an ORDER BY query requires a corresponding composite index. Which change to the indexing policy will allow this query to execute?

Accepted Answer

Add a composite index defined as [{ path: "/status", order: "ascending" }, { path: "/createdDate", order: "descending" }]

Answer

Add two separate range indexes, one on /status and one on /createdDate, since the default indexing policy already indexes every property

Answer

Add a composite index defined as [{ path: "/createdDate", order: "descending" }, { path: "/status", order: "ascending" }]

Answer

Add a spatial index on /status and /createdDate to support multi-field ordering

Question 21

A retail analytics team runs a frequent query against a Cosmos DB container that returns the newest, highest-value orders: SELECT * FROM c WHERE c.status = 'shipped' ORDER BY c.orderDate DESC, c.totalAmount DESC. The query currently fails at runtime with an error indicating the ORDER BY clause cannot be served. The container uses the default indexing policy. What is the most appropriate change to make this multi-property ORDER BY query succeed with minimal RU overhead?

Accepted Answer

Add a composite index on (orderDate DESC, totalAmount DESC) to the indexing policy.

Answer

Add range indexes on both /orderDate/? and /totalAmount/? individually as included paths.

Answer

Enable the wildcard included path /* so all properties are indexed for sorting.

Answer

Switch the container's indexing mode to lazy so ORDER BY across multiple properties is supported.

Question 22

A social media application writes a sequence of user actions to a single-region Cosmos DB account: a user posts a comment (write 1), then edits it (write 2), then deletes it (write 3). The team wants readers in the same region to never see these writes out of order (for example, seeing the delete before the edit), but they want the lowest possible read latency and RU cost that still guarantees ordered reads. They do NOT require reads to always return the most recent write. Which consistency level should they configure?

Accepted Answer

Consistent prefix

Answer

Strong

Answer

Eventual

Answer

Bounded staleness

Question 23

An e-commerce reporting service queries a Cosmos DB container with millions of order documents. The query returns thousands of matching results, but the frontend displays only 50 items per page. Currently the code executes the query, materializes the entire result set into memory, and slices out 50 items per request — consuming large amounts of RUs and memory on every page load. You must minimize RU cost and memory usage while preserving the ability to fetch subsequent pages. What should you do?

Accepted Answer

Set MaxItemCount to 50 on the query request options and use the continuation token returned by each FeedResponse to fetch the next page.

Answer

Add an OFFSET 50 LIMIT 50 clause that increases the offset for each page and re-run the full query per page.

Answer

Enable the integrated cache so the full result set is served from the dedicated gateway on every page request.

Answer

Create a composite index on the queried properties so the entire result set is returned with fewer RUs.

Question 24

A security team wants operations engineers to be able to create and modify Azure Cosmos DB accounts, adjust throughput settings, and manage failover priorities through the Azure portal and ARM, but they must be prevented from reading or writing any document data inside containers. Which approach satisfies these requirements with the least privilege?

Accepted Answer

Assign the engineers a control-plane Azure RBAC role such as Cosmos DB Operator, and do not grant any Cosmos DB data-plane role assignments.

Answer

Assign the engineers the Cosmos DB Built-in Data Contributor data-plane role scoped to the account.

Answer

Distribute the account's primary read-write key to the engineers so they can manage the account.

Answer

Assign the engineers the Owner role on the resource group containing the Cosmos DB account.

Question 25

You maintain a Cosmos DB for NoSQL container named 'Orders' partitioned on '/customerId'. A reporting dashboard runs the query: SELECT * FROM c WHERE c.status = 'Shipped' AND c.region = 'EU'. Neither 'status' nor 'region' is the partition key, and the container spans 40 physical partitions. Users complain the query consumes excessive RUs. Which explanation and remediation is most accurate?

Accepted Answer

The query is a cross-partition (fan-out) query that must be evaluated on all physical partitions; adding a composite index on (status, region) or restructuring to include customerId in the filter reduces the cost.

Answer

The query is a point read that is inefficient because status is not indexed; enabling the default indexing policy will convert it to a single-partition lookup.

Answer

The high cost is caused by the gateway connection mode; switching to direct mode eliminates the fan-out entirely.

Answer

The query already targets a single partition because status equals a constant; the RU cost is unrelated to partitioning and can only be lowered by increasing provisioned throughput.