chnxyzz / clash-exporter Goto Github PK

This is an exporter for Clash, for used by the Prometheus to monitor clash network traffic.

Go to https://github.com/zxh326/clash-exporter/releases download latest release binary file

➜  ./clash-exporter -h
Usage of ./clash-exporter:
  -collectDest
        enable collector dest
        Warning: if collector destination enabled, will generate a large number of metrics, which may put a lot of pressure on Prometheus. (default true)
  -collectTracing
        enable collector tracing.
        It must be the Clash premium version, and the profile.tracing must be enabled in the Clash configuration file. (default false)
  -port int
        port to listen on (default 2112)

git clone https://github.com/zxh326/clash-exporter

# check docker-compose.yml and update environment
cat docker-compose.yml
docker compose up -d

visit localhost:3000 and import example dashboard or via id 18530

tips: grafana default username / password is admin/admin

- job_name: "clash"
  metrics_path: /metrics
  scrape_interval: 1s
  static_configs:
    - targets: ["127.0.0.1:2112"]

groups:
  - name: discard_destination
    rules:
      - record: source_policy_type:clash_network_traffic_bytes_total:sum
        expr: sum without (destination, job) (clash_network_traffic_bytes_total)

• You can import clash-dashboard.json to obtain the example effect, or you can create one yourself based on the following metrics introduction.

• or Import via grafana.com with id 18530

• tracing metrics is empty

  • Required clash premium version
  • Follow clash profile docs enable profile tracing
  • Add -collectTracing=true flag in clash-exporter start script

• high Prometheus Memory

  This may be caused by the default enable of collector destination traffic, which can generate a large number of metrics. Try use -collectDest=false disable it.

• dns query metrics
• proxy dial metrics