The SLOGERT Framework for Automated Log Knowledge Graph Construction

Log files are a vital source of information for keeping systems running and healthy. However, analyzing raw log data, i.e., textual records system events, typically involves tedious searching inspecting clues, as well tracing correlating them across sources. Existing management solutions ease this process with efficient data collection, storage, normalization mechanisms, but identifying linking entities sources enriching background knowledge is largely an unresolved challenge. To facilitate knowledge-based approach to analysis, paper introduces SLOGERT, flexible framework workflow automated construction graphs from arbitrary messages. At its core, it automatically identifies rich RDF graph modelling patterns represent types events extracted parameters that appear in stream. We present the workflow, developed vocabularies integration, our prototypical implementation. demonstrate viability approach, we conduct performance analysis illustrate application on large public dataset security domain.