// the find
hhftechnology/traefik-log-dashboard
A real-time dashboard for analyzing Traefik logs in UI with IP geolocation, status code analysis, and service metrics. Built with React (Shadcn UI) and Node.js.
A self-hosted observability dashboard for Traefik reverse proxy logs, split into a Go agent (15MB, parses logs and exposes a REST API) and a React/Node dashboard (151MB, handles visualization). Aimed at homelab operators and small teams running Traefik who want something more useful than raw log files without setting up a full ELK stack.
The agent/dashboard split is the right call — the Go agent is tiny and does one thing well (tail logs, parse, serve), while the heavier dashboard runs separately. Multi-agent support means you can watch several Traefik instances from one UI, which is genuinely useful for homelab setups with multiple hosts. The GeoIP resolver chain (local MMDB first, HTTP fallback) is practical — works out of the box but lets you plug in a local MaxMind database to avoid rate limits in production. Test coverage in the Go agent is reasonable: there are fuzz tests on the log parser, which is exactly where you'd want them.
The dashboard image is 151MB for what is essentially a Vite SPA plus a Node proxy — that's bloated and mostly traceable to not separating the build toolchain from the runtime image. The GeoIP default uses `ipwho.is`, a free third-party API with no SLA; a high-traffic Traefik instance will hit rate limits quickly and the docs mention this only as a footnote. No authentication on the dashboard itself — the agent has a token, but the dashboard UI is wide open to anyone who can reach port 3000, which is a problem if you're exposing this beyond localhost. The 500/1000 log cap in memory means you lose historical context on busy instances unless you're polling constantly; there's no persistent query layer or time-range search.