Serverless resource management in Amazon Keyspaces (for Apache Cassandra)
Amazon Keyspaces (for Apache Cassandra) is serverless. Instead of deploying, managing, and maintaining storage and compute resources for your workload through nodes in a cluster, Amazon Keyspaces allocates storage and read/write throughput resources directly to tables.
This chapter covers key aspects of resource management in Amazon Keyspaces.
Storage – Amazon Keyspaces provisions storage automatically based on the data stored in your tables. It scales storage up and down as you write, update, and delete data, and you pay only for the storage you use. Data is replicated across multiple Availability Zones
for high availability. Read/Write capacity modes – You can choose between two capacity modes for processing reads and writes on your tables:
On-demand mode (default) – Pay per request for read and write throughput. Amazon Keyspaces can instantly scale capacity up to any previously reached traffic level.
Provisioned mode – Specify the required number of read and write capacity units in advance. This mode helps maintain predictable throughput performance.
Automatic scaling – For provisioned tables, you can enable automatic scaling to adjust throughput capacity automatically based on actual application traffic. Amazon Keyspaces uses target tracking to increase or decrease provisioned capacity, keeping utilization at your specified target.
Burst capacity – Amazon Keyspaces provides burst capacity by reserving a portion of unused throughput for handling spikes in traffic. This flexibility allows occasional bursts of activity beyond your provisioned throughput.
Capacity estimation – This section covers examples of how to estimate read and write capacity consumption for common scenarios like range queries, limit queries, table scans, and lightweight transactions. You can use Amazon CloudWatch to monitor actual capacity utilization. For more information about monitoring with CloudWatch, see Monitoring Amazon Keyspaces with Amazon CloudWatch.