Name some RUs Throughput Modes used in Cosmos DB
In Cosmos DB, there are several throughput modes available for managing request units (RUs) allocation based on the workload patterns and performance requirements of your application. Here are the main RUs throughput modes used in Cosmos DB:
- Standard Provisioned Throughput: This is the traditional mode where you provision a specific amount of RUs per second (RUs/s) for a container or a set of containers. You specify the desired throughput during container creation or modify it later based on your workload needs. This mode provides predictable performance and cost based on the provisioned RUs.
- Autoscale Provisioned Throughput: Autoscale mode is similar to standard provisioned throughput, but it allows Cosmos DB to automatically adjust the provisioned throughput based on the workload demand. It dynamically scales the RUs up or down based on observed usage patterns, ensuring optimal performance while minimizing costs during periods of lower demand.
- Serverless: Cosmos DB introduced a serverless pricing option called “Azure Cosmos DB Serverless” that allows you to pay for the actual resources consumed by your workload. It eliminates the need for upfront provisioning and allows you to scale the database automatically. In serverless mode, you pay per request unit (RU) and storage consumed, and the service automatically manages scaling and performance.
- Free Tier: Cosmos DB offers a free tier with limited resources and capacity that allows you to use Cosmos DB at no cost for low-volume workloads or experimentation. The free tier has certain limitations on throughput and storage capacity but enables you to explore and develop applications without incurring charges.
These throughput modes offer different ways to allocate and manage request units (RUs) in Cosmos DB, providing flexibility to match your application’s requirements and optimize costs. The choice of the throughput mode depends on factors such as expected workload patterns, scalability needs, and budget considerations.