Alibaba Cloud Service Mesh:Description of RateLimitingPolicy fields

Example

The following configuration indicates that throttling is performed on the HTTPBin service in a Kubernetes cluster. The token bucket used for throttling is filled with two tokens every 30 seconds. That is, two requests are allowed every 30 seconds. The maximum number of tokens in the token bucket is two, which indicates that burst traffic is not allowed. Requests are grouped by the user_id request headers. The system determines whether to perform throttling on requests with different user_id request headers separately, and requests with different user_id request headers consume tokens in their own token buckets.

apiVersion: istio.alibabacloud.com/v1
kind: RateLimitingPolicy
metadata:
  name: ratelimit
  namespace: istio-system
spec:
  rate_limiter:
    bucket_capacity: 2
    fill_amount: 2
    parameters:
      interval: 30s
      limit_by_label_key: http.request.header.user_id
    selectors:
    - agent_group: default
      control_point: ingress
      service: httpbin.default.svc.cluster.local

Description of RateLimitingPolicy fields

RateLimitingPolicy defines a global throttling policy in the ASM traffic scheduling suite. It specifies the services on which throttling takes effect, the runtime parameters of the throttling policy, and the attributes of the token bucket used for throttling.

RateLimitingPolicySpec

RateLimitingPolicySpec is the core configuration of RateLimitingPolicy and corresponds to the spec field.

RateLimiter

RateLimiter specifies the parameters of the limiter and the token bucket used by the limiter.

RateLimiterParameters

RateLimiterParameters specify the configuration parameters of the limiter.

RateLimiterRequestParameters

RateLimiterRequestParameters specifies the custom configuration for the limiter to perform operations on requests.

RateLimiterParametersLazySync

RateLimiterParametersLazySync specifies the configuration parameters related to the delayed synchronization capability of the limiter. Delayed synchronization is a capability of the limiter in the ASM traffic scheduling suite. If delayed synchronization is not enabled, Envoy synchronizes with the remote agent for each request to determine whether throttling is required. This allows accurate throttling for a single service. If delayed synchronization is enabled, Envoy makes throttling decisions locally and periodically synchronizes with the remote agent. This reduces the accuracy of throttling but accelerates throttling decisions and reduces request latency.

Description of request labels

In the ASM traffic scheduling suite, request labels are key-value pairs that indicate the characteristics or classification of requests. Scheduling can be performed on different requests based on their labels. For example, requests can be grouped according to specific request labels, and a separate token bucket is used for throttling each group of requests. The labels of a request can be obtained from multiple sources, such as the request metadata and the Baggage header.

Request metadata

The basic metadata of each HTTP request is automatically converted into a series of labels of the request. Mapping between the metadata of an HTTP request and the request labels:

• http.method: the method of the HTTP request. Example: POST.

• http.flavor: the protocol version of the HTTP request. Example: 1.1.

• http.host: the host of the HTTP request. Example: httpbin.default.svc.cluster.local.

• http.target: the path name of the HTTP request. Example: /get.

• http.request_content_length: the size of the HTTP request body. Example: 431.

• http.request.header.header_name: the request header named header_name in the HTTP request headers. For example, http.request.header.user_agent indicates the user_agent request header.

Baggage

Baggage is a standardized mechanism developed by OpenTelemetry to transfer context information across processes in call chains of a distributed system. You can add an HTTP header named baggage to HTTP headers. The value of the baggage header is in the key-value pair format. You can use the baggage header to transfer context data such as tenant ID, trace ID, and security credential. This way, you can use the tracing analysis and log association features without the need to modify code. Example:

baggage: userId=alice,isProduction=false

If a request carries the Baggage header, each key-value pair in the Baggage header is automatically converted into a request label with the same name. For example, the Baggage header in the preceding example is converted into two request labels: userId: alice and isProduction: false.