About The Speaker
I have been using Nagios for 20ish years, and it’s always been my standby for any job I start and have to learn an infrastructure or my homelab. I trust it completely, and I even have it running on a Raspberry Pi under my desk, emailing me when things go wrong.
Leveraging Open Source AI models to predict Outages
The presentation will demonstrate how to leverage data from a Nagios instance, feed it into a model, and easily predict usage trends and capacity planning—all from your laptop in a safe, secure, and local environment.
Behind the Session Title
I spend my time in the AI space. However, I still have to worry about capacity, and if my Nagios machine tells me that my machines are overloaded, wouldn’t it be nice to predict possible issues leveraging the Time Series model I use? I have a small proof of concept (and it should be better as I find more and more time to work on this); my Type 1 diabetes data experimentation inspired this talk, and it’s only natural to start looking at the machines I care about. My goal for this session is to give people a different paradigm of what AI can do with the tools they already have and to be a supporting character for people’s work. tl;dr wouldn’t it be nice to have a 24×7 monitor that can tell you possibly weeks ahead of time that you’re going to have issues so you can plan accordingly? The art of the possible is here with this.
What I Hope You Learn
Ideally my goal is to show a way to leverage our free and Open Source AI Granite model(s) and Nagios technology together so our combined users can find more joy and trust in their systems. The more time they “leave us be,” the more time they can make money and trust us as the experts we are.
Nagios, the Nagios logo, and Nagios graphics are the servicemarks, trademarks, or registered trademarks owned by Nagios Enterprises. All other servicemarks and trademarks are the property of their respective owner. Website Copyright © 2009-2025 Nagios Enterprises, LLC. All rights reserved.
