Performance tuning for Azure Cosmos DB
summary
Explains concrete tuning moves: choose partition key to avoid cross-partition queries (example 202 GB → ~5 physical partitions because of 50 GB / 10,000 RU per physical partition), use change feed to duplicate data into a second container keyed by product ID to avoid costly scans, prefer point reads when you have partition key+id (point read = 1 RU vs query ≈2.3 RU), use direct mode, singleton client and 64-bit SDK, consider dedicated Gateway caching (dedicated gateway caching → 0 RU for 30 min), and a VS Code extension with ~12k downloads to inspect partitions and metrics.
AI-written summary · about
chapters (26)
- 0:00Introduction and agenda
- 1:36Physical partition basics
- 3:26Partition key selection
- 5:18Handling multi-project keys
- 6:30Change feed and data sync
- 8:48Point reads vs queries
- 10:28Connection modes
- 15:07Integrated cache configuration
- 16:04Caching code demonstration
- 17:05Query execution and the SDK
- 18:35Optimizing .NET applications
- 21:18Using the stream API
- 22:18Document client optimization
- 22:40Consistency levels explained
- 23:57Metadata request limits
- 26:06Optimizing bandwidth
- 26:35Partial document updates
- 28:15Patch operations demo
- 29:20Indexing basics and types
- 31:24Composite index rules
- 33:10Composite index examples
- 34:43Modifying indexing policies
- 37:04Throughput and cost savings
- 39:46Priority-based execution
- 40:32Closing thoughts
- 41:02Cosmos DB studio demo
related talks
52mPerformance tuning for Azure Cosmos DB
01h 00mCost cutting strategies for Azure Cosmos DB
08mOptimize for the Best Cost and Performance for Azure Cosmos DB NoSQL API
01h 03mData Modeling and Partitioning in Azure Cosmos DB
14mHow to achieve good cost/performance with Azure Cosmos DB elasticity
32mAzure Cosmos DB Performance Tuning - Sonal Pathak & Amresh Krishanmurthy
33mDriver for Data-Driven Change - Azure Cosmos DB Performance, Security & Cost Optimization Done Right
49mSQL Server vs Azure Cosmos DB