How NVIDIA’s Threat Hunting process leverages Grafana and Loki for log analysis at scale

The security team at NVIDIA, a leading manufacturer of GPU and AI hardware and software, has developed a powerful Threat Detection System (TDS) for analyzing security logs at scale to detect malicious activity. The team relies on Grafana and Grafana Loki to provide a robust platform for analyzing different types of data, such as network logs alongside access logs, to identify and respond quickly to any suspicious activity. This has proven to be an effective and cost-efficient method for identifying threats across several data and log types.

In this session, Senior Software Engineers Amit Singh Hora and Pradeep Thalasta will discuss the current architecture of NVIDIA’s Grafana-Loki stack, which can be deployed on any CSP, including AWS, with Datadog vector as the client for log transmission. The team will share best practices for deploying and managing the stack for scale and multi-tenancy; optimizing performance, efficiency, and cost; and handling unexpected scenarios.

Amit Singh Hora
Senior Software Engineer at NVIDIA
Amit Singh Hora
Senior Software Engineer at NVIDIA
Amit Singh Hora, a Senior Software Engineer at NVIDIA, possesses more than 10 years of experience in developing and designing highly scalable systems capable of handling billions of events. He has a strong passion for open source technologies and has employed and implemented numerous tools during his journey. Additionally, Amit has mentored several bootcamps conducted by different universities. During his leisure time, he enjoys swimming and spending quality time with his family.
Pradeep Thalasta
Senior Software Engineer, Data Science/ML at NVIDIA
Pradeep Thalasta
Senior Software Engineer, Data Science/ML at NVIDIA
Pradeep is a Senior Software Engineer, Data Scientist/ML at NVIDIA working on Cyber AI. Pradeep obtained his Master’s degree in Computer Science from the University of Southern California, LA. He is mainly focused on Deep Learning, Graph Neural Networks and, Distributed GPU computing workflows for data processing and model training. During his leisure time, he enjoys photographing celestial objects.