Skip to main content
Version: 2.4.3

Logging-RocketMQ Plugin

Description#

Apache ShenYu The gateway receives requests from the client, forwards them to the server, and returns the server results to the client. The gateway can record the details of each request,
The list includes: request time, request parameters, request path, response result, response status code, time consumption, upstream IP, exception information waiting.
The Logging-RocketMQ plugin is a plugin that records access logs and sends them to the RocketMQ cluster.

Technical Solutions#

  • Architecture Diagram

  • Full asynchronous collection and delivery of Logging inside the Apache ShenYu gateway

  • Logging platform by consuming the logs in the RocketMQ cluster for repository, and then using Grafana, Kibana or other visualization platform to display

Plugin Usage#

1.Add the Logging-RocketMQ dependency to the gateway's pom.xml file.#

        <!--shenyu logging-rocketmq plugin start-->        <dependency>          <groupId>org.apache.shenyu</groupId>          <artifactId>shenyu-spring-boot-starter-plugin-logging-rocketmq</artifactId>          <version>${project.version}</version>        </dependency>        <!--shenyu logging-rocketmq plugin end-->

2.In shenyu-admin --> Basic Configuration --> Plugin Management --> loggingRocketMQ, configure the rocketMQ parameter and set it to on.#

2.1Open the plugin and configure rocketmq, configure it as follows.#

  • The individual configuration items are described as follows:
config-itemtypedescriptionremarks
topicStringMessage Queue Topicsmust
namesrvAddrStringMessage queue nameserver addressmust
producerGroupStringLog Message Producer Groupmust
sampleRateStringSampling rate, range 0~1, 0: off, 0.01: acquisition 1%, 1: acquisition 100%Optional, default 1, all collection
compressAlgStringCompression algorithm, no compression by default, currently supports LZ4 compressionOptional, no compression by default
maxResponseBodyIngeterMaximum response size, above the threshold no response will be collectedOptional, default 512KB
maxRequestBodyIngeterMaximum request body size, above the threshold no request body will be collectedOptional, default 512KB

Except for topic, namesrvAddr, producerGroup, all others are optional, in most cases only these 3 items need to be configured.

2.2 Configuring Selectors and Rulers#

For detailed configuration of selectors and rules, please refer to: Selector and rule management。 In addition sometimes a large gateway cluster corresponds to multiple applications (business), different applications (business) corresponds to different topics, related to isolation,
then you can configure different topics (optional) and sampling rate (optional) by selector, the meaning of the configuration items as shown in the table above.
The operation is shown below:

Logging information#

The fields of the captured access log are as follows.

Field NameMeaningDescriptionRemarks
clientIpClient IP
timeLocalRequest time string, format: yyyy-MM-dd HH:mm:ss.SSS
methodrequest method (different rpc type is not the same, http class for: get, post wait, rpc class for the interface name)
requestHeaderRequest header (json format)
responseHeaderResponse header (json format)
queryParamsRequest query parameters
requestBodyRequest Body (body of binary type will not be captured)
requestUriRequest uri
responseBodyResponse body
responseContentLengthResponse body size
rpcTyperpc type
statusresponse status
upstreamIpUpstream (program providing the service) IP
upstreamResponseTimeTime taken by the upstream (program providing the service) to respond to the request (ms ms)
userAgentRequested user agent
hostThe requested host
moduleRequested modules
pathThe requested path
traceIdRequested Link Tracking IDNeed to access link tracking plugins, such as skywalking,zipkin

Consumption and display of Logging#

As each logging platform has differences, such as storage available clickhouse, ElasticSearch, etc., visualization has self-developed or open source Grafana, Kibana, etc..
Logging-RocketMQ plugin uses RocketMQ to decouple production and consumption, while outputting logs in json format, consumption and visualization require users to choose different technology stacks to achieve their own situation.

Panel Display#

Users can choose to visualize the implementation according to their own situation.
The following shows the effect of Grafana:
Grafana Sandbox Experience