Azure Cosmos DB Serverless and Autoscaling Capabilities [Build 2020]

May 19, 2020
Sun rising over Earth

Azure Cosmos DB serverless addresses intermittent traffic and “bursty” workloads for small apps

App developers of all kinds want to build apps quickly — at any scale. To make building apps faster and more flexible, Microsoft is announcing offers and features for Azure Cosmos DB, its fast NoSQL database with open APIs for any scale. These additions are designed to provide customers with pricing models to fit their
unique needs, shorten the app development cycle and effectively manage their workloads.

Developers building apps managing smaller numbers of users need to manage intermittent traffic patterns and “bursty” workloads in a cost-effective way. To address these unique needs, Microsoft has designed Azure Cosmos DB serverless, a more cost-effective way to use Azure Cosmos DB.

This new model offers per-operation compute pricing ideal for noncritical, intermittent and small workloads with moderate performance requirements. Consumption-based pricing bills customers for the total resources (request units) used for each discrete database operation.

Customers must create an Azure Cosmos DB account within the Azure portal and select “serverless” as their pricing model. All resources within this account will run in serverless mode and get billed based on actual consumption of resources to deliver each database request.

Visit the Azure blog to learn more.

Azure Cosmos DB autoscale provisioned throughput offers scalability for largescale workloads

Microsoft is also announcing the forthcoming general availability of Azure Cosmos DB autoscale provisioned throughput (previously called autopilot), which offers SLA-backed single-digit millisecond latency and 99.999% availability guaranteed while taking the planning and management of throughput capacity off customers’ hands, making it well-suited for large, unpredictable workloads. After customers select a max level for autoscale provision throughput, Azure Cosmos DB automatically scales provisioned throughput up and down based on the workload usage — relieving customers of the need to manage scale by themselves. All features will be available in Summer 2020.

Visit the Azure blog to learn more.

If your organization is interested in learning more about NoSQL databases, please contact Octavian Technology Group by email at

Cameron Vetter
My name is Cameron Vetter and I'm the Principal Architect at the Octavian Technology Group.
Microsoft MVP. Cloud. Enterprise Architecture. Machine Learning. Mixed Reality.
Transformative and Impactful,
Put Us to Work for You
Contact Us