Hosters, servers, services and domains — discovered automatically and laid out as a tree, level by level. From the cloud account at one end to every domain it serves at the other.
Three concrete uses, all of them improving the day someone breaks the build.
When an incident fires on one node, the graph shows the blast radius — every service that talks to it, with health colour-coded.
The fastest way for a new engineer to learn what your fleet looks like. No outdated Confluence page, no architect explaining to the new hire what the boxes mean.
Replay the graph at any timestamp. Was that connection there before the change? Did it appear after? The map is auditable, not just current.
The graph is generated from observable infrastructure state. You can add manual edges where they help — but you don't have to.
Connect Hetzner, DigitalOcean, Vultr, AWS — your servers come in already grouped under their hoster. No manual entry, no drift.
The agent reads systemd, listening ports, and process tree on each host. nginx, apache, caddy, mysql, postgres, redis, php-fpm — all show up automatically.
Web-server config files are parsed for server_name / ServerName / domain blocks. Each domain is bound to the service that serves it — no domain you have to register by hand.
Because it built itself, and it's still building itself.