Cosmos DB性能问题（即RU费用，吞吐量，Http 429） [英] Cosmos DB Performance Issue (ie. RU Charge, Throughput, Http 429)
问题描述
1) My api performance was getting down continuously which used to connect to cosmos db collection X.
2) Most of the suggestion on internet was that your database should have PatitionKey which should be decided properly like country, city etc.
3) Based on such suggestion, created a separate db with collection 'Y' with partitionkey as DocumentType whose value range from 1 to 30. Converted all the stored procedure, query, and code based on partitionedkey.
4) My concerned is even after using partition key, giving same throughput to new partioned db as earlier one, and on firing same query on both cosmos db don't see any difference in performance.
以下是疑问：
1）想知道什么应该是宇宙数据库的平均吞吐量设置大小：4GB和15个开发人员不断访问它。
1) Wanted to know what should be average throughput set for cosmos db with Size: 4GB and 15 developers accessing it continously.
2）即使在对数据库进行分区之后，在两个cosmos db上执行查询或存储过程时，RU也没有太大差异。
2) Even after partitioning the database dont see much difference in RU when executed query or stored procedure on both cosmos db.
推荐答案
嗨Brijkishor，
Hi Brijkishor,
有三个方面要做Azure Cosmos DB详见以下文章： 性能和规模使用Azure Cosmos DB进行测试 它们是：
There are three aspects to Azure Cosmos DB as detailed in the following article: Performance and scale testing with Azure Cosmos DB and they are:
第一部分是对特定实例进行基准测试：
The first part is to benchmark your specific instance:
随后是客户端调优和服务器端分区：
Followed by the client-side tuning and server-side partitioning:
此外，您可以专门深入了解查询性能： 使用Azure Cosmos DB调优查询性能
Additionally, you can dig into query performance specifically: Tuning query performance with Azure Cosmos DB
我已经看到了查询响应时间现有客户之前每个帖子的1-2秒范围。这是在客户端测量的。
I have seen query response times in the 1-2 second range per previous posts from existing customers. This is measured at the client.
这篇关于Cosmos DB性能问题（即RU费用，吞吐量，Http 429）的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
