TechNuggets Academy

Design data storage solutions

Free Microsoft Certified: Azure Solutions Architect Expert practice — 6 questions on Design data storage solutions, with explanations. No sign-up. Full 12-question mixed test →

Question 1 of 6 · Design data storage solutions
A global retail company uses Azure Cosmos DB (SQL API) with multi-region writes enabled across three regions. The application must guarantee that a user immediately sees their own writes on their next read within the same session, while keeping RU consumption and latency as low as possible. Which consistency level BEST meets these requirements?
Session consistency guarantees read-your-own-writes and monotonic reads within the client session while still allowing multi-region writes, at far lower latency and RU cost than stronger consistency levels — ideal for user-facing write scenarios in a multi-master account.
Question 2 of 6 · Design data storage solutions
An IoT platform ingests telemetry from 50 million devices into a Cosmos DB container. A small percentage of devices are far more active than the rest, and queries frequently filter by device and by date. The current partition key of deviceId is causing hot partitions on the busiest devices. Which partition key strategy BEST resolves this while keeping queries efficient?
Combining deviceId and date creates a high-cardinality synthetic key that spreads writes across many logical partitions, even for the busiest devices, while still enabling efficient queries filtered by device and date without cross-partition fan-out.
Question 3 of 6 · Design data storage solutions
A financial services firm must retain trade records for 7 years in Azure Blob Storage in a way that even a Storage Account Owner cannot alter or delete during that period, satisfying regulator audit requirements. During active litigation, some records also need an indefinite legal hold applied without disturbing the retention timers on unrelated blobs. Which configuration BEST meets these requirements?
Version-level immutability policies provide WORM (write once, read many) protection that no principal, including an Owner, can override during the retention period, and legal holds can be applied per version independently of the time-based policy on other blobs — this satisfies compliant, auditable immutability.
Question 4 of 6 · Design data storage solutions
A company is migrating an on-premises SQL Server 2016 workload that heavily relies on SQL Server Agent jobs, cross-database transactions, linked servers, and CLR integration. The target platform must offer near-100% engine compatibility, built-in automatic high availability without manually configuring Availability Groups, no OS patching responsibility, and in-memory OLTP performance. Which target BEST meets all requirements?
Managed Instance provides near-100% SQL Server engine compatibility, including SQL Agent, cross-database transactions, linked servers, and CLR, and the Business Critical tier includes built-in automatic HA (three synchronous replicas plus in-memory OLTP) with no manual Availability Group configuration or OS patching required.
Question 5 of 6 · Design data storage solutions
An application running on Azure Database for PostgreSQL Flexible Server must survive a complete failure of an entire Azure availability zone with automatic failover and near-zero RPO, and the standby replica must reside in a different zone than the primary. Which configuration BEST satisfies this requirement?
Zone-redundant HA deploys the standby replica in a different availability zone from the primary, using synchronous replication with automatic failover, providing protection against a full zone/datacenter failure with near-zero RPO.
Question 6 of 6 · Design data storage solutions
A media company stores read-heavy static assets in an Azure Storage account. The design must tolerate the loss of an entire datacenter within the primary region via synchronous replication, and must also allow the application to read data from a secondary region immediately during a regional outage, without waiting for a failover to complete. Which redundancy option should be selected?
RA-GZRS combines synchronous zone-redundant replication within the primary region (surviving a datacenter/zone failure) with asynchronous geo-replication to a secondary region, and the 'RA' capability grants read access to the secondary region's data at any time, even before a failover occurs.
Ready for the real thing?

The full course: two full-length practice tests, video lessons for every exam domain, hands-on labs and detailed explanations.

undefined $34.99 with code 00651DDA8F57721374B8 — valid through Sep 23.

Get my $34.99 deal →