Go Monitoring

This tool helps to monitor network services (at the moment HTTP, Ping and raw TCP)
It exposes Prometheus timeseries that can be used by Alertmanager for notification. This is inspired by Blackbox Exporter.

Usage

Docker:

docker run -p 8080:8080 -v $PWD/config.yaml:/config/config.yaml ghcr.io/the-kube-way/go-monitoring:latest

Kubernetes: Check deploy/kubernetes/manifests.yaml.

Exported Prometheus timeseries

In addition to standard Go metrics, Go monitoring exports the following timeseries:

• go_monitoring_up: 1 if target is up, else 0
  • probe: name of the probe (http, ping, raw_tcp)
  • id: id of the target (url for http probe, host for ping, host:port for raw tcp)
  • name: name of the target (as specified in the config, "" if not specified)

Configuration

The configuration is done via a YAML file:

global:
  check_interval: 5m  # Default value if not overwritten per target
  oncall_offer: "day" # Possible values: day,we,24_7. Default to day if not defined.

http:
  - url: https://example.com
    name: monitor_example_com
    expected_status_code: 200
  
  - url: https://example.com/search
    name: monitor_example_com_search
    check_interval: 30s
    method: POST
    body: '{"test": "test"}'
    skip_verify: false
    headers:
      Authorization: Basic xxx
      Content-Type: application/json
    expected_status_code: 200
    expected_response_body: "OK"

ping:
  - host: example.com
    name: example_server
    check_interval: 1m

raw_tcp:
  - host: ssh.example.com
    port: 22
    name: ssh_example_com
    check_interval: 10m

License

MIT