A node group is a Redis replica group containing one primary node and one or more replica nodes. The primary accepts reads and writes; replicas accept reads. section 11.6.2
Replication copies data from the primary to replicas, distributing read traffic and increasing read throughput. Replicas synchronize asynchronously, so they eventually reach the primary's state. Redis clients must understand the cluster topology and select the appropriate node for each command. section 11.6.2
Use a node group when choosing how a Redis deployment should scale:
- Choose replication when one node cannot handle the request volume and traffic is mostly reads; it also increases availability at no extra cost. section 11.6.2
- Choose sharding when the dataset or workload exceeds one node's capacity; each shard stores part of the key space and can contain its own replicas. section 11.6.2
- Use Application Auto Scaling to increase or decrease the number of shards automatically from utilization metrics. section 11.6.2
