Redis Quota

The redisquota adapter can be used to support Istio’s quota management system. It depends on a Redis server to store quota values.

This adapter supports the quota template.


redisquota adapter supports the rate limit quota using either fixed or rolling window algorithm. And it is using Redis as a shared data storage.

Example configuration:

redisServerUrl: localhost:6379
connectionPoolSize: 10
  - name: requestCount.quota.istio-system
    maxAmount: 50
    validDuration: 60s
    bucketDuration: 1s
    rateLimitAlgorithm: ROLLING_WINDOW
      - dimensions:
          destination: ratings
          source: reviews
        maxAmount: 12
      - dimensions:
          destination: reviews
        maxAmount: 5
Field Type Description
quotas Params.Quota[]

The set of known quotas. At least one quota configuration is required

redisServerUrl string

Redis connection string : ex) localhost:6379

connectionPoolSize int64

Maximum number of idle connections to redis Default is 10 connections per every CPU as reported by runtime.NumCPU.


Field Type Description
dimensions map<string, string>

The specific dimensions for which this override applies. String representation of instance dimensions is used to check against configured dimensions. dimensions should not be empty

maxAmount int64

The upper limit for this quota override. This value should be bigger than 0


Field Type Description
name string

The name of the quota

maxAmount int64

The upper limit for this quota. max_amount should be bigger than 0

validDuration google.protobuf.Duration

The amount of time allocated quota remains valid before it is automatically released. This is only meaningful for rate limit quotas. value should be 0 < valid_duration

bucketDuration google.protobuf.Duration

bucketduration will be ignored if ratelimitalgorithm is FIXEDWINDOW value should be 0 < bucketduration < validduration

rateLimitAlgorithm Params.QuotaAlgorithm

Quota management algorithm. The default value is FIXED_WINDOW

overrides Params.Override[]

Overrides associated with this quota. The first matching override is applied.


Algorithms for rate-limiting:

Name Description

FIXED_WINDOW The fixed window approach can allow 2x peak specified rate, whereas the rolling-window doesn’t.


ROLLING_WINDOW The rolling window algorithm’s additional precision comes at the cost of increased redis resource usage.