Azure SQL Database is Microsoft's fully managed PaaS relational database service offering near-infinite scale with Hyperscale architecture supporting databases up to 100TB, intelligent performance optimization through AI-driven automatic tuning and Intelligent Query Processing, and enterprise-grade security with built-in features like Always Encrypted and ledger verification. Azure SQL Managed Instance extends this with near-100% SQL Server compatibility, providing instance-level features like SQL Agent, cross-database queries, and CLR integration—making it the optimal lift-and-shift migration target for on-premises SQL Server workloads moving to the cloud. A key mental model: SQL Database excels for cloud-native greenfield applications requiring elastic scale and serverless compute, while Managed Instance serves as a compatibility bridge preserving existing SQL Server investments and enabling hybrid cloud scenarios with the Managed Instance Link for near real-time replication.
What This Cheat Sheet Covers
This topic spans 16 focused tables and 72 indexed concepts. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.
Table 1: Service Tiers and Purchasing Models
Picking a tier and a purchasing model is the first—and most consequential—decision you make, because it sets the ceiling on scale, latency, and price for everything that follows. The vCore tiers (General Purpose, Business Critical, Hyperscale) trade off remote versus local storage and the number of replicas, while the choice between DTU and vCore billing decides whether you get a simple bundled metric or the granular control that unlocks Hybrid Benefit, reserved capacity, and savings plans.
| Tier | Example | Description |
|---|---|---|
4 vCores, 20 GB storageserverless: 0.5-2 vCores | • Budget-balanced tier with remote storage (Azure Premium Storage) separated from compute • supports serverless auto-pause/resume and scales 1-128 vCores • availability SLA: 99.99% with single replica | |
8 vCores, zone redundantread scale-out: 3 replicas | • Mission-critical tier with local SSD storage for lowest latency • includes 3 secondary replicas (Always On availability groups) with one readable for offloading queries • availability SLA: 99.995% (zone redundant: 99.995%). | |
up to 100 TB database4 read replicas, rapid scale | • Distributed architecture with tiered storage (page servers + log service) enabling databases up to 100TB • rapid compute scaling (~2 mins regardless of data size) and up to 4 named read replicas for horizontal scale-out • fast backups (file-snapshot based). |