Select Language:
When ElastiCache automatically adds new shards to your cluster in a replication group with auto scaling enabled, it takes care of placing the primary and replica nodes across different availability zones (AZs). The system evenly distributes these nodes within your subnet group to keep your data safe and ensure your application stays available, even if one zone experiences issues.
However, there’s a catch. Currently, you can’t specify which AZs new nodes should go to when auto scaling adds capacity. For example, if you want a primary node in us-east-1c and replicas in other specific AZs, auto scaling won’t let you make that choice. It only adds nodes based on overall resource needs, like CPU usage or scheduled times, and doesn’t give you control over exactly where the nodes are placed.
This creates a gap when compared to manual scaling, where you can directly choose the AZs for your nodes. Having the option to control AZ placement during auto scaling would give you more flexibility and help you meet specific needs related to latency, compliance, or high availability tied to certain zones.
While this isn’t currently available, requesting this feature could lead to more precise architectural control without sacrificing the benefits of auto scaling. It would make managing complex workloads easier and more effective, especially in setups that need to follow strict zoning rules to function optimally.
ChatGPT
Perplexity
Gemini AI
Grok AI
