guardus

Feature-equivalent v1 reimplementation of Gatus backend on top of plumego stable roots and extensions.

This use-case sits alongside use-cases/gatus (the original Gatus drop-in) to validate plumego in monitoring/observability product use cases.

Status

v1 backend complete: HTTP/TCP/ICMP/DNS/TLS/STARTTLS/gRPC/WebSocket probes, ExternalEndpoint push, SQLite + memory storage, Basic Auth, the five v1 alert providers (custom/email/slack/telegram/discord), maintenance windows, connectivity checker, Prometheus metrics, /health, and embedded Vue SPA. See AGENTS.md for v1 scope.

Build

make build
make run

Configuration

Deployment settings come from environment variables (GUARDUS_*). See env.example for the full surface — copy it, edit values, and source (or set the variables in your orchestrator) before make run.

Endpoint definitions live in the storage backend (SQLite by default at ./guardus.db). On each startup, if GUARDUS_BOOTSTRAP_FILE points at an existing JSON document, every entry is upserted by key into the store, then the watchdog runs against whatever the store now holds:

{
  "endpoints": [
    {
      "name": "front-end",
      "url": "https://example.org",
      "interval": "5m",
      "conditions": ["[STATUS] == 200"]
    }
  ]
}

The bootstrap file is optional — once endpoints are in the database they persist across restarts. An admin API for managing endpoints at runtime is the next milestone; for now, edits made directly to the DB take effect after restart.

Frontend

The SPA is rewritten in React + Vite (source in web/app/). Run:

make frontend-build   # installs Node deps and builds into web/static/
make build            # Go binary embeds web/static/ via //go:embed

The dev server proxies /api, /health, /metrics to http://localhost:8080 so npm run dev works against a running backend.

Suites and OIDC are intentionally omitted (out of v1 scope).