Anomaly Detection and Prediction for often Non Normally Distributed data

I have been using TICK stack for a while and I love it!
Thank you guys for coming up with such useful tools.

I have a question
We are monitoring server where data isn’t often normally distributed and where metrics such as CPU and Memory usage and disk may or may not have a constant pattern or seasonality.
What are my available options and what should be the best approach for anomaly detection/alerting and Prediction?

I started reading an article by Anais, it will take some time, but hoping if someone out there already gone ahead can give a direction.