As I understand the problem, the difference in command counts is due to ElastiCache being a managed service. AWS restricts certain high-privilege commands (like SHUTDOWN, REPLICAOF, or certain CONFIG subcommands) that could interfere with the managed infrastructure, which is why they are excluded from the @admin ACL category.

If you're attempting to change database parameters (which typically fall under the admin category), ElastiCache doesn't allow this directly on the node via a Redis client. Instead, you need to create a custom Cache Parameter Group in the AWS Console or via CloudFormation, Terraform, or AWS CLI. You can modify almost all variables there that Redis OSS allows via CONFIG SET, and the service will safely roll out these changes to your nodes.

In short, for a managed service like ElastiCache, administrative changes should be performed via the AWS Management Console, CLI, or Infrastructure as Code, rather than by sending raw Redis commands via a client.

Serverless caches:

• https://docs.aws.amazon.com/AmazonElastiCache/latest/dg/RedisConfiguration.html#RedisConfiguration.Serverless

Node-based clusters:

• https://docs.aws.amazon.com/AmazonElastiCache/latest/dg/ParameterGroups.Engine.html#ParameterGroups.Redis