Explore a comprehensive 43-minute conference talk on implementing practical AI with machine learning for observability in Netdata. Delve into the traditional monitoring pipeline and its limitations before discovering the Netdata approach to distributed monitoring. Learn about the challenges and benefits of AI in observability, including how machine learning models work and are shared. Examine Netdata's implementation of anomaly detection, scoring engines, and the Anomaly Advisor feature. Gain insights into the accuracy and trustworthiness of AI-driven anomaly detection, and understand how it can enhance monitoring capabilities. Conclude with a look at the future of machine learning in Netdata and its potential impact on observability practices.
Practical AI with Machine Learning for Observability in Netdata
Overview
Syllabus
intro
preamble
about netdata
how it is different - the traditional way
the traditional monitoring pipeline
the result of the traditional pipeline
the netdata way
how it works
distributed monitoring
fully on-prem
distributed mixed
the benefits
netdata vs prometheus
ai for observability
ai for observability is tricky
how ml works
sharing of ml models
so what it can do?
is anomaly detection accurate?
then, how can we trust it?
how it can help us?
what does netdata do with ml?
anomaly rate on every chart
a netdata chart
netdata's scoring engine
a netdata dashboard - what is anomalous?
anomaly advisor
highlights of ml in netdata
what is next for ml in netdata?
thank you!
Taught by
Conf42
