Skip to main content

Monitor

This chapter introduces Go-Sail's monitoring.

Introduction​

For running application services, service monitoring is of critical importance; it renders the operational status of applications more transparent and clear, while also contributing significantly to performance optimization and—in a microservices environment—elastic scaling.

Currently, Go-Sail utilizes two approaches—Prometheus and PProf—to implement telemetry and observability.

Metrics​

Go-Sail utilizes Prometheus to monitor various service metrics. You can collect this data via the metrics endpoint for subsequent use. The following metrics are currently being collected:

  • HTTP Request Latency
  • HTTP Status Code Counts
  • CPU Utilization
  • Memory Utilization
  • Disk Utilization
  • Network Traffic

Configuration​

If you need to enable metric monitoring, configuration is required. We have excerpted the specific sections that need to be configured.

config.example.yaml
http_conf:
  debug: false
  addr: ""
  swagger_conf:
    enable: false
    redoc_ui_path: ""
    json_path: ""
    favicon_path: ""
  prometheus_conf:
    enable: true
    addr: ""
    access_path: ""
    disable_system_sample: false
    disk_path: ""
    sample_interval: ""

Collection​

Once the configuration is complete, Go-Sail will enable Prometheus monitoring. However, please note that this monitoring does not actively push data; instead, it exposes an access_path for external access, allowing you to collect metrics using tools such as Grafana.

PProf​

Configuration​

PProf configuration is tied to the HTTP configuration; simply set the debug parameter within the HTTP configuration section to true.The access URL is /debug/pprof.

config.example.yaml
http_conf:
  debug: true
  addr: ""
  swagger_conf:
    enable: false
    redoc_ui_path: ""
    json_path: ""
    favicon_path: ""
  prometheus_conf:
    enable: false
    addr: ""
    access_path: ""
    disable_system_sample: false
    disk_path: ""
    sample_interval: ""

Usage​

PProf supports viewing the following:

  • memory usage
    /debug/pprof/heap

  • CPU profiling
    /debug/pprof/profile

  • goroutine status
    /debug/pprof/goroutine

  • blocking analysis
    /debug/pprof/block

  • lock contention
    /debug/pprof/mutex

You can use the go tool command to perform a detailed analysis. For example:

go tool pprof http://localhost:8080/debug/pprof/heap