Files
serverAppConfigs/serverAppConfigs-docs/Infrastructure/Homepage.md
T
Frank 80a070d71b homepage: three cards could never be green, and one of them was lying
The board showed mariadb, valkey and Postgres as "down". None of them is down:
the checker runs on Dankclaw and can reach that host's loopback, its container
names, and published LAN addresses — while all three cards probe
container-internal database ports (192.168.0.235:3306, 127.0.0.1:6379/:5432).
They were red by construction, which is how a status board stops being read.
Publishing a database port on the LAN to make a badge green is not a trade this
fleet makes, so they are wip with the reason in their description, and the apps
that front them carry the status. Gluetun (WireGuard, UDP) was the pre-existing
example of the same call.

The rule is now written down in the Homepage doc: a card must be probeable from
the monitor's host or be wip. That is the same lesson as the Prometheus jobs
that scrape HTML pages and can never report up — an always-red signal teaches
the operator to ignore the surface it lives on.
2026-09-16 07:36:11 -07:00

221 lines
11 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# Homepage
A password-protected web site on Dankclaw lists the services that are up, with a short description of each and a link to both the service and its usage documentation.
**Served by:** Caddy, `file_server` + `forward_auth` (Authentik SSO) — **Docs:** this page
## Status — deployed to production (2026-08-23), on the naked domain since 2026-08-24
The homepage is live on Dankclaw. `./scripts/deploy.sh dankclaw` copies the static assets to `/srv/homepage/`, installs the status checker (systemd timer, every 3 min, writes `status.json`), writes `deployed.js` (the last master deployment datetime), renders the Caddyfile (the shared basicauth hash from the host `.env`), reloads Caddy, and checks health. The Caddyfile auth uses a `route` block with `forward_auth` (see [Auth](#auth)); the caddy container mounts the host's `/srv/homepage` (not the repo source) so the generated files serve.
**Moved to the naked domain (2026-08-24):** the homepage now serves `franksdank.cloud` and `www.franksdank.cloud`. The old names `dankclaw.franksdank.cloud` and `dankserver.franksdank.cloud` redirect to it. Authentication moved from per-page basicauth to Authentik `forward_auth` (see [Authentik](Authentik.md)).
The per-service `/docs/` pages, the *arr stack, and the optional `/how/` page remain separate threads.
## Design decision — 2026-08-22
The live up/down check runs server-side. A checker on Dankclaw probes each backend and writes a same-origin `status.json`. The page fetches only that file and renders the badges. The browser never calls a backend directly. The checker runs as a systemd timer every 3 minutes.
The site has two pages. The root lists user-facing services only, with no host labels. An `/admin/` subpage lists the admin and infrastructure services behind a separate password for the operator. Gitea is a user service. Hermes is an admin service. A "how it is built" page for visitors is deferred (Q6).
The client-side probe design is blocked in this topology. The page origin is `https://franksdank.cloud`. The service URLs are `https://<name>.franksdank.cloud`. That is cross-origin and auth-gated. Both fail a browser `fetch`.
Intended direction: move to client-side probing. Serve every service as a same-origin `https` path under `franksdank.cloud`. Add a health endpoint and CORS headers to each backend. Then the page probes each service directly. Keep this note until that shift is complete.
## Structure
The site has two pages. The root page lists the user services. The `/admin/` page lists the admin and infrastructure services. The optional `/how/` page is deferred (Q6).
```mermaid
graph TD
subgraph site["franksdank.cloud + www (Caddy)"]
ROOT["/ user services<br/>(shared password)"]
ADMIN["/admin/ admin services<br/>(operator password)"]
HOW["/how/ how it is built<br/>(deferred)"]
STATUS["/status.json"]
end
TIMER["systemd timer<br/>every 3 minutes"]
CHECKER["checker script<br/>probes each backend"]
BACKENDS["backends on both hosts"]
TIMER --> CHECKER
CHECKER -->|writes| STATUS
STATUS -->|polled| ROOT
STATUS -->|polled| ADMIN
CHECKER -->|probes| BACKENDS
```
## Status model
Each service shows one of three states.
```mermaid
stateDiagram-v2
[*] --> Up
[*] --> Down
[*] --> WIP
Up --> Down: probe fails
Down --> Up: probe succeeds
WIP --> Up: work completes
```
- **Up** — the probe reached the backend.
- **Down** — the probe failed (connection refused or timeout).
- **WIP** — a static override. The service is intentionally not finished. The checker does not probe it.
## Checker
The checker runs on Dankclaw as a systemd timer. Every run probes each backend directly, then writes `status.json`. It probes the backend `host:port`, not the Caddy site. A Caddy `502` or `401` cannot tell up from down. The probe method is per service (`tcp` or `http`). The timeout is 5 s. The page reads `status.json` once on load. It does not auto-refresh.
```mermaid
flowchart TD
A[timer fires] --> B[read service config]
B --> C{for each service}
C -->|wip| D[mark WIP]
C -->|probe| E[probe: tcp or http]
E -->|connected| F[mark Up]
E -->|refused / timeout| G[mark Down]
D --> H[write status.json]
F --> H
G --> H
```
## Service classification
**User services** (root page): SearXNG, ComfyUI, OpenWebUI, Gitea, Calibre-Web, RomM, Copyparty, Plex, Jellyfin.
**Admin services** (`/admin/`): Hermes, Grafana, Prometheus, Loki, Caddy, Pi-hole, Gluetun, the LLM fleet, and the low-level backends (mariadb, redis, Postgres, node-exporter, alloy). The backends show status only. They get a link only where a UI is meaningful.
**Excluded from the listing:** the *arr stack + qBittorrent (Prowlarr, Sonarr,
Radarr, Lidarr, Readarr, qB) — live on FrankBook-Pro but operator-only: no
vhosts, direct IP:port access only. See [[FrankBook-Pro]].
## Layout
```
shared/homepage/
├── index.html # root page (user services)
├── admin.html # admin + infrastructure services
├── how.html # deferred (Q6)
├── style.css # full theme; a toggle strips to bare HTML
├── app.js # loads status.json once, renders the panels
├── services.json # service config (user + admin)
└── fonts/ # Terminus (woff2) for @font-face
```
`status.json` is generated at runtime in `/srv/homepage/`. It is not committed.
## Rendering
The page renders a **grid of TUI-style panels** — one panel per service. Each panel shows the state badge, the name, the description, and the `Open` and `Docs` links. A summary line at the top counts the states. The footer shows the last master deployment datetime (Q43).
The page reads `status.json` **once on load**. It does not auto-refresh (Q41).
Styling is **toggleable** (Q27). The default is the full terminal theme. A toggle strips the page to bare, unstyled HTML (motherfuckingwebsite.com style). The page does not alter the user's cursor (Q28).
## Auth
Caddy wraps the site with `forward_auth` (Authentik SSO). A single Authentik
session covers the whole site: the user page, the `/admin/` page, and the
shared data + assets. The browser's Authentik cookie is reused for the data
fetches, so no per-page credential is needed. See [Authentik](Authentik.md) for
the provider, the outpost, and the first-time setup.
The Caddyfile pattern (one of the forward_auth vhosts):
```
franksdank.cloud www.franksdank.cloud {
root * /srv/homepage
route {
# 1. The outpost path is proxied to Authentik (login UI + session).
reverse_proxy /outpost.goauthentik.io/* authentik-server:9000
# 2. forward_auth checks the session before serving the page.
forward_auth authentik-server:9000 {
uri /outpost.goauthentik.io/auth/caddy
copy_headers X-Authentik-Username X-Authentik-Groups X-Authentik-Email X-Authentik-Name
trusted_proxies private_ranges
}
# 3. The pages, in file order. 'handle' preserves order; a top-level
# 'rewrite' would be sorted first and break the path matchers.
handle /admin* {
try_files /admin.html
file_server
}
@data path /services.json /status.json /app.js /style.css /deployed.js
handle @data {
file_server
}
handle {
try_files /index.html
file_server
}
}
}
```
A `handle` inline matcher is a single token; the multi-path `@data` match is a
named matcher. The caddy container mounts the host's `/srv/homepage` (where the
deploy writes the static files and the checker writes `status.json`), not the
repo's `shared/homepage`. A compose change needs a container recreate (`podman
compose up -d caddy`), not a reload. The Caddyfile syntax is documented in the
agent knowledge base (`~/.pi/agent/docs/caddy/caddyfile.md`).
## Service entry format
```json
{
"name": "SearXNG",
"category": "user",
"description": "Privacy-respecting metasearch engine.",
"url": "http://search.franksdank.cloud",
"docs": "/docs/searxng/",
"probe": "127.0.0.1:8081",
"method": "http",
"wip": false,
"showLink": true
}
```
- `category` — `user` or `admin`. Selects the page.
- `url` — the "Open" link (the Caddy site).
- `docs` — the per-service docs page (`/docs/<svc>/`).
- `probe` — the backend `host:port` the checker probes.
- `method` — `tcp` or `http`. The probe type for this service (Q10).
- `wip` — when true, the badge shows WIP and the checker skips the probe.
**A card must be probeable from the monitor's host, or it must be `wip`.** The
checker runs on Dankclaw, so its reach is: that host's own loopback, its own
container names, and LAN addresses that are actually *published*. A card whose
probe is a container-internal port of another host can never turn green — it
reports a red dot for a healthy service, which is how the board stops being
worth reading (the same lesson as the Prometheus jobs that scrape HTML pages).
Three cards were exactly that until 2026-09-16 — `mariadb`
(`192.168.0.235:3306`), `valkey` and `Postgres` (`127.0.0.1:6379`/`:5432`) — all
probing unpublished database ports. Publishing a database on the LAN to make a
badge green is not a trade this fleet makes, so they are `wip` with the reason in
their description, and the app that fronts each database carries the status
(Glutun, a WireGuard UDP service, was the pre-existing example of the same
call).
- `showLink` — when false, the panel shows status only (no "Open" link).
## Deployment and testing
The deploy is **git-driven**. Master always reflects the currently deployed state (Q33). `deploy.sh` copies the assets to `/srv/homepage/`, installs the checker timer, renders the Caddyfile (both hashes), and reloads Caddy. It stamps the deployment datetime the footer shows (Q43).
The build is **full TDD** (Q36–Q38). Tests exist before the code: the checker (a fake up port, a down port, a WIP service, and the JSON schema), the pages and config (the `services.json` schema, `caddy validate`, the `status.json` reference), and the auth check (the forward_auth block and the renderable basicauth placeholder).
## Per-service docs (separate workstream)
Each service gets a docs page at `/docs/<svc>/` (Q30). The content is a **how-to for a regular user** — no maintenance details (Q29). The pages are password-protected by the same auth as the homepage section they belong to (Q31). They are built as each service goes active (Q32). The `docs` link on a service points to its page. The existing reference docs in `serverAppConfigs-docs/` are not linked from the homepage. They may contain secrets (Q22).
## See Also
- [[Homepage Design Grilling]] — the decision log
- [[Caddy]] — the proxy that serves it
- [[Dankclaw Stack]], [[Dankserver Stack]] — the services it lists