Introduction

Argux is a monitoring framework that is designed to work slightly different from most monitoring applications that you’ll find on the web. It is designed with the following aspects in mind.

Note

Argux is still under heavy development, some of the statements below might still be a bit ‘optimistic’. (eg. they will be true once there is a release) - These remarks also count as reminders to what it is that is being developed.

Decentralisation

In short, most of the time you are not interested in ping times from your monitoring host to server X. You are interested in ping times from the offices in Paris and Madrid to the server in Berlin.

REST API

It is impossible to prepare for any scenario of hosts or services that require monitoring. Therefor it is important that the interface to Argux is clear and easy to use. It will have client API’s in several programming languages, but if all else fails you could control it with CURL.

The REST API is the center of the application, it is meant for use by external scripts but it is also used internally and as backend to the Web-interface.

Encryption

Since Argux is using basic HTTP requests, you can simply configure your favourite reverse-proxy or SSL terminator to run all traffic via HTTPS instead of plain HTTP.

Authorisation

All requests to Argux require authentication. A user is allowed to write metrics of one host, or many. Or is only allowed to read information.

Example 1:

Basically, I only trust the user ‘postgres’ on the database-server to store any posgres related information.

Example 2:

An agent running on server X is not able to save metrics about server Y, unless is is explicitly allowed.