Skip to main content

Monitors

Purpose​

A monitor is an external healthcheck which is performed on your application. For now we're implementing only http healthcheck with GET, POST and PUT methods. In the future we'll have more healthchecks.

Our monitor system is able to:

  • present the last check in our graphical interface
  • export the result as a Prometheus/OpenMetrics gauge in the /v1/metrics endpoint of CWCloud's API
  • push the metrics to an opentelemetry OTLP/grpc endpoint (configured with the OTEL_COLLECTOR_ENDPOINT environment variables, see the selfhosted documentation for that).

Demo​

Here's a 10 min demo that will show you how to use our monitor management system:

Creating a monitor​

Observability > Monitor

monitors_create_interface

Notes:

  • For the expected http code, you can use the * meta caracter which is evaluated as any number. For example if you set 20*, the monitor will be "success" with every http code starting by 20 like 200, 201, 202...
  • The response body contain will make the monitor fail if the response body doesn't contain the value
  • You can setup a basic authentication and custom headers
  • For the POST and PUT methods you can configure the request's body
  • You can also use the cli to create monitors, see this documentation

Indexing Prometheus metrics into Quickwit​

  • In order to install Quickwit via CWCloud, go read this tutorial
  • In order to configure Vector and Quickwit to scrap and index the Prometheus metrics, see this blogpost