EvoParser: An Evolutionary Approach to Log Parsing

In the domain of network operations and manage-ment, efficient log parsing is crucial for processing the highvolumes of log data necessary to monitor the reliability andperformance of the operated infrastructures and services. Thispaper introduces a novel method for creating highly efficientlog parsers. Our method employs a genetic algorithm to evolveoptimized log parsing graphs that deliver high-quality parsingresults. Experimental evaluations demonstrate that our approach achieves results comparable to or better than state-of-the-art techniques while maintaining low complexity during parsing.These characteristics make it particularly suitable for real-timeapplications in high-volume data scenarios.