Rate Limiting Advanced plugin allowing for limits much larger than specified using ElastiCache cluster in AWS

When using the Rate Limiting Advanced plugin with an ElastiCache cluster in AWS we are specifying a limit of 10 however we are passing 15-20 transactions before receiving the error “429 - Too many requests”.

{
	"message": "API rate limit exceeded"
}

For example we have a limit of 10 and a window size of 30. About 20 of these requests will pass inside that 30 second window.

This type of behavior can be experienced if the Cluster is being specified as a host inside the Rate Limiting Advanced plugin.

The reason this occurs is because it is creating multiple keys for storage allowing for more than the specified limit. Under redis-cli we can run the keys command to confirm:

keys *
1) "1660576260:30:abHx1lXb5BTdBjfke34BlNInAdajk350L"
2) "1660576290:30:abHx1lXb5BTdBjfke34BlNInAdajk350L"

To correct this, we would need to grab the Cluster hostname from AWS and specify it under the value Config.Redis.Cluster Addresses.

Note: You will need to remove the port from Config.Redis.Port and add it to the cluster_hostname specified under Config.Redis.Cluster Addresses, like the example above.

Now while testing we will have the exact limit hit during the window size that is configured on the Rate Limiting Advanced plugin.

Again to confirm this, we can connect to our redis-cli and run the keys command:

keys *
1) "1660576830:30:abHx1lXb5BTdBjfke34BlNInAdajk350L"

This time we only see 1 key being stored.