What's changed: Created DP-420 Chapter 3 (Domains 2+3: replication & consistency (five levels strong/bounded staleness/session/consistent prefix/eventual; freshness vs RU/latency), automatic/manual failover, multi-region writes, conflict resolution (LWW/custom); analytical integration (analytical store/Azure Synapse Link HTAP/Microsoft Fabric mirroring vs Spark connector/CDC/Change Feed + Azure Functions denormalize-referential-aggregate-archive/Event Hubs/Azure AI Search)).
3.2Analytical workloads and service integration
Understand the analytical store and Azure Synapse Link, Microsoft Fabric mirroring vs the Spark connector, Change Feed with Azure Functions for denormalization/referential integrity/aggregation/archiving, and Event Hubs and Azure AI Search integration.
To flow Cosmos DB transactional data into analytics or other services without impacting operations, the analytical store and Change Feed are central.
3.2.1Analytical store, Synapse Link, and Fabric
Enabling the analytical store on a container auto-syncs transactional data into a columnar analytical copy (HTAP). Azure Synapse Link lets Synapse Spark/SQL query the analytical store directly, doing large-scale analytics without impacting the transactional store (operational RU). Additionally, Microsoft Fabric mirroring for Cosmos DB mirrors data into Fabric OneLake near-real-time for analysis in Fabric (distinct from the batch-oriented Cosmos DB Spark connector). The analytical store also supports Change Data Capture (CDC).
3.2.2Change Feed and service integration
The Change Feed lets you read container changes (create/update) in time order. With Change Feed + Azure Functions you implement, event-driven: updating denormalized copies, enforcing referential integrity, maintaining aggregations (reporting), and archiving data. You can also stream events to other apps via Azure Event Hubs and index into Azure AI Search for full-text/vector search. Distinguish "react to changes to drive other processing" = Change Feed from "large-scale analytics without operational impact" = analytical store/Synapse Link.
Cues: "large-scale analytics without operational RU impact" = analytical store + Synapse Link (HTAP). "near-real-time analysis in Fabric" = Fabric mirroring (batch = Spark connector). "react to changes for denormalize/referential/aggregate/archive" = Change Feed + Azure Functions. add search = AI Search.
Watch the mix-ups: (1) Analytical store (no-impact analytical copy) vs Change Feed (change-event stream). (2) The analytical store must be explicitly enabled on the container. (3) Synapse Link requires the analytical store. (4) Fabric mirroring (near-real-time) vs Spark connector (batch) differ in purpose.
3.2.3Section summary
- Analytical store + Synapse Link = no-impact HTAP analytics; Fabric mirroring for near-real-time analysis
- Change Feed + Azure Functions = event-driven denormalize/referential/aggregate/archive
- Event Hubs for event integration; Azure AI Search for full-text/vector search
Sign in to track progress — Log in.
Quick check
(just a quick review)Q1. You want to run large-scale analytics on Cosmos DB data without impacting transactional operational RU. Best?
Q2. You want event-driven updates of denormalized copies and maintained aggregations in reaction to container changes. Best?
Q3. You want to mirror Cosmos DB data into Microsoft Fabric OneLake near-real-time for analysis. Best?
Q4. Which correctly distinguishes the analytical store from Change Feed?
Q5. You want to add full-text/vector search over Cosmos DB data. Best integration target?