TechPrufer
Get started
IBM
IBM logoIBM · Feature trail

Log Template Mining for Network Outage Detection

1 post · May 2020

What this trail covers

The Drain3 log-template mining project, an open-source streaming log template miner, has been refactored for Python 3 and enhanced for production use. It now includes improved streaming support, resiliency through Kafka-based state persistence, and masking for improved accuracy. The project is available as a PyPI package and on GitHub. Log templates can be used for analytics, detecting anomalies in message frequencies and parameter values to identify network outages and other issues.

All IBM trails
  1. Use open source Drain3 log-template mining project to monitor for network outagesThis post details the refactoring of the original Drain log-template mining tool into Python 3 (Drain3), including improvements for streaming support and resiliency by using Kafka for state persistence. It also introduces masking for improved accuracy and packaging as a PyPI package. The post explains how log templates can be used to create time-series data for anomaly detection, enabling early identification of network outages.