Microsoft Azure Cosmos DB Developer Specialty
225問の練習問題
最終確認:April 2026
学習のための個人ノートとリソースリンク
認定でフィルター
DP-420 is Microsoft's specialty credential for developers building applications on Azure Cosmos DB for NoSQL. It validates the ability to design and implement data models, plan partition and distribution strategy, integrate Cosmos DB with surrounding Azure services, optimize performance and cost, and maintain Cosmos DB solutions in production. The audience is professional developers and data engineers who write Python, .NET, or JavaScript / TypeScript against the Cosmos DB SDK. The exam is heavily code-and-modeling focused: expect 40–60 questions in 100 minutes including code-completion drag-and-drops (SDK snippets, SQL API queries), scenario items, and at least one case study.
Largest domain at 37%. Document modeling for NoSQL workloads (denormalization, embedding vs. referencing), partition-key design, change-feed patterns, indexing policies (included / excluded paths, composite indexes, spatial indexes), and TTL configuration.
About 8%. Multi-region replication, multi-region writes, consistency-level tradeoffs (strong / bounded staleness / session / consistent prefix / eventual), conflict-resolution policies, and global distribution patterns.
About 8%. Change feed processor, Azure Functions Cosmos DB triggers, Event Hubs / Kafka integration, Cosmos DB analytical store with Azure Synapse Link, and integration with Azure AI Search.
About 17%. Request unit (RU) sizing and tuning, autoscale vs. manual throughput, indexing optimization, query performance, and cost analysis with the capacity calculator.
About 30%. Backup and restore (continuous and periodic), disaster recovery, security (Microsoft Entra auth, RBAC, customer-managed keys, IP firewall, Private Endpoint), monitoring (Azure Monitor, diagnostic logs), and SDK retry / error handling.
$110k–$150k–$210k USD annual
Range covers US-based mid-to-senior backend developers where Cosmos DB proficiency is required. Senior engineers building globally distributed apps at FAANG / fintech often clear $230k TC. Cosmos-DB-specific salary data is sparser than for general Azure roles given the narrower talent pool; figures lean on adjacent NoSQL / cloud-developer roles.
Source: levels.fyi 2025 backend / cloud developer roles, U.S. BLS OEWS May 2024 (15-1252 software developers, 15-1242 database administrators), Glassdoor 2025. Figures are approximate; actual compensation depends on role, region, and experience.
DP-420 sits in a narrower but well-paid niche — applications that genuinely need globally distributed, low-latency, multi-API NoSQL storage. Demand is concentrated at gaming companies, IoT platforms, retail / e-commerce at scale, and Microsoft-partner consultancies. Recruiters use it as a strong signal of deep Cosmos DB modeling and tuning competence, which commands a premium given the limited pool of qualified candidates. It pairs naturally with AZ-204 (Developer Associate) for full-stack Cosmos developers and with AI-102 / AI engineering roles where Cosmos DB serves vector and operational data for RAG architectures. Demand has been steady, with modest growth from Cosmos DB's expansion as a vector store for GenAI applications through 2024–2026.
There are no formal prerequisites. Microsoft recommends practitioner-level developer experience (one to two years professional development) plus working familiarity with Cosmos DB. Candidates without prior Cosmos DB exposure typically need significant additional time. AZ-900 and DP-900 are useful conceptual on-ramps for candidates new to Azure or to NoSQL data platforms; AZ-204 is highly complementary, given DP-420 assumes Azure-developer-level fluency with SDK patterns, Microsoft Entra authentication, and managed identities.
Proficiency in C#, Python, or JavaScript / TypeScript is essentially required: code-completion drag-and-drops show real Cosmos DB SDK snippets, with .NET examples most heavily represented in Microsoft's study material. The official Microsoft Learn path covers all five domains in roughly 30–40 hours. Hands-on time is essentially required — a personal Azure subscription with a small Cosmos DB account (or the Cosmos DB free tier) lets candidates practice partition-key design, indexing policies, and change-feed scenarios.
DP-420 sits in the Specialty tier and is broadly considered moderately to highly challenging — comparable to AZ-204 in code-completion difficulty, with narrower but deeper Cosmos-DB-specific surface area. Plan on 70–110 hours of study over 8–12 weeks for candidates with prior Cosmos DB experience; substantially longer otherwise. The exam runs about 100 minutes with 40–60 questions in multiple-choice, multiple-response, drag-and-drop (including code-completion), hot-area, and case-study formats. Case studies are timed separately and cannot be revisited.
The most common stumbling block is partition-key selection — the exam consistently presents nuanced workload patterns and expects candidates to identify the partition key that distributes load evenly while keeping common queries on a single partition. Indexing policy questions (included / excluded paths, composite indexes, query cost analysis) are another consistent trap area. As a Specialty exam, third-party study material is sparser; lean primarily on Microsoft Learn and the Cosmos DB documentation.
Most recent skills-measured update. Added vector-search coverage for AI workloads, expanded continuous-backup framing, modernized Microsoft Entra and customer-managed-key content. Microsoft refreshes DP-420 less frequently than role-based exams given its specialty status — typically every 18–24 months.
Restructured into the current five-domain layout, expanded change-feed and Synapse Link coverage, and integrated continuous-backup content.
Initial GA as Microsoft's first dedicated Cosmos DB developer credential. Original outline focused on the SQL (Core) API only and emphasized partitioning, RU sizing, and SDK patterns.
DP-420 (Microsoft Azure Cosmos DB Developer Specialty) is a a deeply specialized exam covering advanced topics in a narrow domain — expect hands-on experience to be a prerequisite Specialty-level exam. Most candidates need 100–200 hours of study spread over 2–4 months for specialty exams. These assume hands-on experience in the specialty domain. Most candidates who score consistently above the passing threshold on practice exams pass on their first attempt.