Horus: Causal Analysis of Distributed Systems Logs

Horus is a system that enables the refinement of distributed system logs in a causally-consistent fashion. This repository lists all parts of Horus and their repositories.

Deployment configuration of Horus pipeline using Docker.

Log4j Appender that forwards log entries of Java applications to Horus pipeline.

The eBPF tracer used in Horus. It is based on the state-of-the-art Falcon's tracer.

Visualizer for Horus causal graphs. This project is based on ShiViz, a visualization engine that renders interactive communication graphs based on logs augmented with vector clock timestamps.

Neo4j plugin includes procedures to perform efficient causal queries on causal graphs generated by Horus.

Collection of serializers to serialize events coming from Horus tracer to the Horus processor module. The serializers follow the FlatBuffers specification placed in Horus Tracer repository.

Processor module in the Horus pipeline. Combines both processors of intra- and inter-process causal relationships.

Benchmark scripts for testing parts of Horus