How does the alternative to Blue/green deployment work

Amazon OpenSearch Service has introduced new capabilities that allow updating both node count and volume configurations without requiring traditional blue/green deployments. Here's how it works:

1. For clusters without dedicated cluster manager nodes:

• You can directly update the number of data nodes
• You can make changes to cluster volume size, volume type, IOPS, and throughput

This new approach offers several advantages:

• Significantly reduces deployment time
• Minimizes resource utilization compared to blue/green deployments
• Allows updates with minimal performance impact to online traffic

Important limitations to note:

• For volume updates, this is only available for domains with current generation instances and when volume size per data node is within 3 TiB
• Decreasing storage volume size or making storage volume changes within short intervals will still require a blue/green deployment
• If you modify both data node instance type and count simultaneously, a blue/green deployment will still be required

This solution does reduce downtime compared to traditional blue/green deployments because it eliminates the need to maintain two parallel environments during updates. With the traditional blue/green approach, you needed to create and maintain a complete replica environment during the transition, which required more resources and time.

You can verify whether your specific change requires a blue/green deployment by using the dry-run option before making the changes.

This feature is available for domains running any OpenSearch version, and Elasticsearch versions 7.1 and above, across all AWS Regions where Amazon OpenSearch Service is available.

• https://aws.amazon.com/about-aws/whats-new/2024/02/amazon-opensearch-service-update-cluster-volume-without-blue-green/
• https://aws.amazon.com/about-aws/whats-new/2024/04/amazon-opensearch-service-update-node-count-without-blue-green/