配置故障注入

混沌工程对于构建可靠系统至关重要。fault-injection 插件可以在请求到达上游服务之前，在网关层模拟故障和延迟，从而测试应用的韧性。

前置条件

• API7 企业版实例正在运行。
• 已创建网关组且网关实例正在运行。
• 按照从控制台获取令牌中的步骤准备一个令牌。

配置 HTTP 中止

可以对一定比例的请求返回指定 HTTP 状态码，以模拟服务故障。

Admin API

curl -k "https://localhost:7443/apisix/admin/services/fault-injection-service?gateway_group_id={gateway_group_id}" -X PUT \
  -H "X-API-KEY: ${API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "fault-injection-service",
    "upstream": {
      "type": "roundrobin",
      "scheme": "http",
      "nodes": [
        {
          "host": "httpbin.org",
          "port": 80,
          "weight": 100
        }
      ]
    }
  }'

curl -k "https://localhost:7443/apisix/admin/routes/fault-injection-route?gateway_group_id={gateway_group_id}" -X PUT \
  -H "X-API-KEY: ${API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "fault-injection-route",
    "paths": ["/get"],
    "service_id": "fault-injection-service",
    "plugins": {
      "fault-injection": {
        "abort": {
          "http_status": 503,
          "body": "Service Temporarily Unavailable",
          "percentage": 10
        }
      }
    }
  }'

ADC

adc.yaml

services:
  - name: fault-injection-service
    upstream:
      nodes:
        - host: httpbin.org
          port: 80
          weight: 1
    routes:
      - name: fault-injection-route
        uris:
          - /get
        plugins:
          fault-injection:
            abort:
              http_status: 503
              body: "Service Temporarily Unavailable"
              percentage: 10

adc sync -f adc.yaml

配置响应延迟

还可以延迟响应，以模拟缓慢的网络状况或后端延迟。

Admin API

curl -k "https://localhost:7443/apisix/admin/services/delay-fault-injection-service?gateway_group_id={gateway_group_id}" -X PUT \
  -H "X-API-KEY: ${API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "delay-fault-injection-service",
    "upstream": {
      "type": "roundrobin",
      "scheme": "http",
      "nodes": [
        {
          "host": "httpbin.org",
          "port": 80,
          "weight": 100
        }
      ]
    }
  }'

curl -k "https://localhost:7443/apisix/admin/routes/delay-fault-injection-route?gateway_group_id={gateway_group_id}" -X PUT \
  -H "X-API-KEY: ${API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "delay-fault-injection-route",
    "paths": ["/delay/5"],
    "service_id": "delay-fault-injection-service",
    "plugins": {
      "fault-injection": {
        "delay": {
          "duration": 5,
          "percentage": 20
        }
      }
    }
  }'

ADC

adc.yaml

services:
  - name: delay-fault-injection-service
    upstream:
      nodes:
        - host: httpbin.org
          port: 80
          weight: 1
    routes:
      - name: delay-fault-injection-route
        uris:
          - /delay/5
        plugins:
          fault-injection:
            delay:
              duration: 5
              percentage: 20

adc sync -f adc.yaml

验证配置

要验证注入比例为 10% 的中止配置，请发送多个请求：

for i in {1..20}; do curl -i "http://127.0.0.1:9080/get"; done

应有部分请求收到注入的 503 Service Temporarily Unavailable 响应。对于延迟，当故障被注入时，curl 命令应等待配置的 duration 时长。

要验证延迟示例，请测量对延迟路由发出的几个请求：

time curl -i "http://127.0.0.1:9080/delay/5"

如果应用配置后路由立即返回 404，请等待几秒，让最新配置下发到网关后重试。

后续步骤

• 配置代理镜像——使用真实生产流量测试应用。
• 配置响应改写——在向客户端返回响应前修改响应数据。
• 配置限流——保护系统免受流量突增影响。