Debugging and optimizing Azure Cosmos DB performance | Azure Friday

Поделиться
HTML-код
  • Опубликовано: 24 ноя 2024
  • Deborah Chen joins Scott Hanselman to share some best practices on how to debug and optimize Azure Cosmos DB for better performance. Watch as they go through the common issues newcomers to Azure Cosmos DB run into with respect to performance and how to solve them by tuning Request Unit (RU) cost and choosing a good partition key.
    Azure Cosmos DB - Globally distributed, multi-model database service for any scale
    aka.ms/azfr/56...
    Optimize provisioned throughput cost in Azure Cosmos DB
    aka.ms/azfr/56...
    Optimize query cost in Azure Cosmos DB
    aka.ms/azfr/56...
    Partitioning in Azure Cosmos DB
    aka.ms/azfr/56...
    Request Units in Azure Cosmos DB
    aka.ms/azfr/56...
    Create a free account (Azure)
    aka.ms/azfr/56...
    #Microsoft #MicrosoftAzure #CosmosDB

Комментарии • 15

  • @bharathb4803
    @bharathb4803 5 лет назад +5

    Thank you so much guys. You really helped me solve some of my product issues.

  • @vivekgarg185
    @vivekgarg185 3 года назад

    Excellent showcase Deborah and Scott. Thanks

  • @mcbacker
    @mcbacker 5 лет назад +1

    Great video with a ton of tips in a small amount of time. Thank you Deborah and Scott. 👍

  • @gert-janvanderkamp3508
    @gert-janvanderkamp3508 3 года назад

    Excellent vid! Thanks guys and get well soon Scott!

  • @gert-janvanderkamp3508
    @gert-janvanderkamp3508 3 года назад

    For query performance consideration I would defenitely consider partion key as an index. The clustered index basically, with id as the second step. If you can navigate those to steps that's the fastest (and cheapest) way.

  • @demite5267
    @demite5267 4 года назад

    Really great demo, Deborah you're awesome!

  • @rasmustchristensen
    @rasmustchristensen 5 лет назад

    Great video. In the last demo you concated your own Partitionkey. Is it possible to use it by parts? Lets say your store some Hotels. They are partitioned by the city their in. Then you want to get all the newest hotels overall. This would as I see it make a cross partition query, is it possible to avoid that?

    • @MicrosoftAzure
      @MicrosoftAzure  5 лет назад

      Thanks for reaching out! To get the most recent N items without a cross-partition query, see the approach in this article: msft.social/Jsl2Ez . It involves denormalizing the data into another container.
      For cross partition queries, if your workload data size is small - e.g.

  • @mathewjamesn
    @mathewjamesn 5 лет назад

    Deborah - your presentation looks really good and I shall definitely taken care these tips for my Cosmos db database as well. One quick question - Are the samples you have shown uploaded in github ? It will be great if you can share those. Thanks!

    • @mcbacker
      @mcbacker 5 лет назад +2

      Here it is github.com/deborahc/cosmos-perf-tips ✌️

    • @mathewjamesn
      @mathewjamesn 5 лет назад +1

      @@mcbacker Thank You!

  • @davidntumba2447
    @davidntumba2447 2 года назад

    "If you do nothing you could scale infinitely"

  • @andrecoleman1856
    @andrecoleman1856 2 года назад

    Lol thamks money

  • @SlavenOlujic
    @SlavenOlujic 2 года назад

    Deborah is like Microsoft's female version of Kanye West