!NADA – !Not Adaptive Thresholds

Project Notes

0 - What is !NADA?
^^^^^^^^^^^^^^

!NADA is a brand new project which intents to insert baseline adaptive thresholds to Nagios(R) monitoring framework. By "adaptive" I mean a threshold which may change through time, accordingly to a given resource behavior.

This project is intended those ones, who have already did a question like:
"How can I avoid false positives when monitoring a given server that every Monday has a higher load average than during the other days?"

1 - How does it work?
^^^^^^^^^^^^^^^^^

!NADA requires a MySQL or a SQLite installation together with Nagios(R). It encapsulates your check plugin, parses and stores performance data into DB, calculates the standard deviation and creates two new metrics, pointing to the top and bottom of your baseline. If collected value overflow (up or down) the
baseline, !NADA change the plugin return code to CRITICAL thus causing Nagios(R) to alert.

!NADA' standard behavior assumes that you are using a week seasonality, if it's not appropriate, please may the source be with you.

Let's explain how it works by a simple example:

If a given check occurs just now (let's say: Monday at 11:07 PM), !NADA will retrieve the last one hundred Monday =~ 11:07 PM check results from DB. It will then calculate the standard deviation using these one hundred check results and make a good baseline to current check.