Posted On: May 22, 2024
Starting today, Amazon Redshift is making snapshot isolation as the default for provisioned clusters when you create a new cluster or restore a cluster from a snapshot. The database isolation level will remain unchanged on your existing provisioned clusters unless explicitly changed. You can switch to serializable at any time if it is your preferred database isolation level. This change makes the product experience consistent for both Provisioned and Serverless which already uses snapshot isolation as default.
Amazon Redshift offers two database isolation levels — serializable and snapshot — to handle concurrent transactions within your data warehouse. Serializable isolation provides strict correctness guarantees equivalent to running your operations serially. Most data warehousing applications do not need these strict guarantees that limit concurrency on operations. Unlike serializable, snapshot isolation gives you better performance by allowing for more concurrency of operations on the same table when processing large volumes of data. You can change the isolation level for your database using CREATE DATABASE or ALTER DATABASE commands.
Snapshot isolation will be enabled by default for provisioned clusters in Amazon Web Services China (Beijing) Region, operated by Sinnet, and the Amazon Web Services China (Ningxia) Region, operated by NWCD.
To learn more about how to use Redshift database isolation levels, please refer to the Amazon Redshift documentation.