Simple HTTP monitor

A Simple HTTP monitor written in Golang that exposes its findings to Prometheus

The monitor reads its target from env vars, and exposes a Summary with quantiles: 0.5, 0.9, 0.99

SETUP

1. The following env vars can be set for configuring the monitor

Running the monitor without setting variables will Monitor This project's docker hub page once every 10 seconds and expose the metrics on port 9100

• scrapePort - The port on which Prometheus will scrape the metrics (defautl 9100)
• subsystem - The subsystem (i.e. server/website/store - defualt website)
• monitorUrl - The URL that should be monitored (default https://hub.docker.com/repository/docker/tomgurdev/simple-go-http-monitor)
• monitorInterval - how often the monitor should run in seconds (defautlt 10)
• componentName - the name of the monitored service (default simple-http-monitor-docker-hub)

2. Run the monitor (3 options)
   • After exporting all env vars, issue the following command go run monitor.go
   • After exporting all env vars, build the go executable and run it go build monitor.go; ./monitor
   • Build a docker image containing the monitor app and run it docker build -t monitor:1 .; docker run --name monitor --restart always -d -p 9100:9100 -e "scrapePort=9100" -e "subsystem=website" -e "monitorUrl=https://hub.docker.com/repository/docker/tomgurdev/simple-go-http-monitor" -e "monitorInterval=10" -e "componentName=simple_http_monitor_docker_hub" mymonitor:1

Metrics

• Metrics will be exposed at the host IP on the port set in scrapePort at the /metrics endpoint (i.e. 1.2.3.4:9100/metrics)
• The exposed metrics are:
  • The round-trip duration in seconds in 3 quantiles: 0.5, 0.9, 0.99
  • The HTTP response status

from Label

When run on an AWS EC2 instance, a label will be added to all metrics from=<availability zone>.

Whne run outside of AWS, the from label will be filled with the outgoing IP address of the server.