Summary

Homer offers a 100% open-source observability solution for SIP, VoIP, and RTC traffic, consolidating packet capture, storage, and analysis into an all-in-one application. Built with a modern Golang codebase and a data lake architecture, it simplifies the deployment and operation of robust monitoring systems for network engineers and VoIP administrators.

How It Works

Homer 11.x is a Golang monolith for X64/ARM64 Linux/macOS, featuring a data lake powered by DuckDB 1.5 and Apache Arrow/IPC/Parquet. Its modular design includes Ingest (HEP reception), Storage (DuckLake catalog, Parquet), Node (gRPC/HTTP, FlightSQL), and Coordinator (REST API). This approach enables columnar OTLP, on-demand querying, linear scaling over object storage, flexible schemas, and backward compatibility with HEPv3 agents.

Quick Start & Requirements

Build with make. Run the all-in-one server via ./homer --config-path /etc/homer/homer.json. Requires a Go development environment. Configuration is JSON-based, with an interactive wizard (homer wizard) generating profiles (e.g., all-in-one, writer, edge). Documentation and examples are available in the repo.

Highlighted Details

• All-in-one application integrating ingest, storage, node, and coordinator.
• Data lake architecture using DuckDB/Arrow for columnar OTLP and efficient querying.
• Cloud-native design for K8s and standard deployments, scalable down to zero.
• MCP support and LLM/Agent-friendly API for automation.

Maintenance & Community

No specific details on maintainers, sponsorships, or community channels were found in the provided README.

Licensing & Compatibility

Released under the AGPL-3.0 License. This strong copyleft license requires derivative works to also be AGPL-3.0, potentially restricting integration with proprietary or closed-source software.

Limitations & Caveats

The AGPL-3.0 license may limit commercial use cases requiring closed-source integration. While designed for scalability, the monolithic nature might require careful management in very large distributed setups, despite claims of linear scaling over object storage.