ws: migrate push stream to data streaming

- daemon: send full snapshot on connect; versions/tick now carry the
  full state of one subsystem (subsystem + data); no legacy
  updated/subsystems payloads; refresh_state and POST /status/refresh
  broadcast per-subsystem versions with data
- client: modelSet() patches models in place; onMessage/topic refresh
  retired; 3s initial-load fallback via new POST /api/status/refresh
- schema: lib/schema.py TypedDicts + hoover/schema.js defaults +
  docs/state-model.md as single source of truth for state shapes
- system: poll at 1s, volatile metrics registered, dashboard uses a
  dedicated system model (status model removed)
- firewall: refuse to strip both https and ssh from the default zone
  (409, force override via UI confirm); set_zone_services persists
  services to the declarative config; collector exposes default_zone
- UI: pages migrate to flat state shapes; post-mutation modelFetch
  refreshes removed (WS delta covers it)
- tests: ws snapshot/delta/broadcast, refresh-state, schema types,
  model-set/js ws handler and reconnect fallback
This commit is contained in:
2026-08-20 01:38:00 +00:00
parent 9c9f92ad04
commit 332d14e37d
45 changed files with 2819 additions and 496 deletions
+12 -9
View File
@@ -148,19 +148,22 @@ The daemon runs background polling tasks for subsystems with external runtime st
| wireguard | 10s | Peer connections/handshakes change frequently |
| dnsmasq | 10s | Lease file + service status |
| networkd | 10s | Interface up/down, DHCP address changes |
| system | 1s | Real-time metrics (load/memory/swap/traffic) |
| nginx | 60s | Config-file drift self-heal (lazy in-place migration) |
| acme | 300s | Config-file drift self-heal (lazy in-place migration) |
nginx, acme, and auth are not polled — they have no external runtime state.
Only `auth` is not polled — it has no external runtime state.
**Two-layer diff:** Each poll cycle classifies changes as:
- **Structural change** (zones added, peers removed, config changed): triggers `bump()` + broadcast `{"type": "versions", ...}` → full UI re-load
- **Volatile change only** (transfer counters, DHCP-assigned IPs): sends `{"type": "tick", "subsystems": [...]}` → lightweight per-subsystem re-fetch
- **Structural change** (zones added, peers removed, config changed): triggers `bump()` + broadcast `{"type": "versions", "subsystem": ..., "data": ...}` → daemon pushes the full subsystem data over WS; the client patches the model in place via `modelSet`
- **Volatile change only** (transfer counters, DHCP-assigned IPs): sends `{"type": "tick", "subsystem": ..., "data": ...}` → same in-place patch, without a version bump
- **No change**: silence
Volatile fields per subsystem: `wireguard` (peer transfer/handshake stats), `firewall` (DHCP-assigned IPs), `networkd` (DHCP addresses, link metrics). Defined per collector via `register_volatile()`.
Volatile fields per subsystem: `system` (load/memory/swap/traffic), `wireguard` (peer transfer/handshake stats), `firewall` (DHCP-assigned IPs), `networkd` (DHCP addresses, link metrics). Defined per collector via `register_volatile()`.
Poll intervals are configurable via `VACUUM_WALL_POLL_INTERVALS` env var (`firewall:30,wireguard:10,...`).
On collector failure during a poll, no broadcast is sent (avoids noisy ticks). State data is set to `None`.
On collector failure during a poll, no broadcast is sent (avoids noisy ticks). State data is set to `None`, and `broadcast_versions` additionally skips a `None` payload defensively (a null payload would clobber good client data — the next successful poll or mutation broadcasts the real value).
## System Config Import
@@ -338,11 +341,11 @@ Client loads /static/app.js ──→ Hoover initializes, checkSession() (401 wi
Authenticated ──→ mounts #sidebar and #main render roots
apiFetch() ──→ injects Authorization: Bearer <token> header ──→ Flask REST API
Flask before_request ──→ validates JWT from header, checks blacklist, verifies permissions
Hoover connects WebSocket ──→ daemon/ws (127.0.0.1:9091?token=<access_token>)
Hoover connects WebSocket ──→ daemon/ws (raw JWT as Sec-WebSocket-Protocol subprotocol name; legacy `Bearer <token>` subprotocol + X-Auth-Token header fallbacks accepted)
Page navigate (hash change) ──→ reactive router state updates ──→ render engine re-executes ──→ VDOM diff patches DOM
User action (form submit) ──→ apiFetch() ──→ Flask REST API ──→ daemon/client.py ──→ vacuum-walld
Token expiry ──→ refreshScheduler() ──→ POST /api/auth/refresh ──→ new tokens
WebSocket message (versions) ──→ topic match ──→ page load() re-executed ──→ state updated ──→ render engine patches DOM
WS connect ──→ snapshot (full state) / versions + tick deltas (per-subsystem data) ──→ modelSet() patches model in place ──→ render engine VDOM-diffs and patches only changed DOM nodes
```
The SPA entry point only serves `index.html` at `/`. All other paths return 404. Non-API, non-static paths are not served by Flask — the client-side router handles all navigation via hash changes. A dedicated `/vendor/<path>` route serves vendored JS libraries.
@@ -355,9 +358,9 @@ Each route is a `definePage()` component with reactive state, async data loading
All JavaScript is served as ES modules. Cache invalidation is handled via HTTP cache-control headers. Dev mode (`VACUUM_WALL_DEV`) disables aggressive static asset caching.
### WebSocket Broadcast
### WebSocket Data Streaming
The daemon broadcasts state-change notifications via WebSocket. Hoover's `subscribe` mechanism maps page-level topic subscriptions to automatic `load()` re-executions. Messages are debounced (300ms) and in-flight loads are aborted before re-loading, ensuring the UI always displays the latest available data.
The daemon pushes state over the WebSocket — no HTTP round-trip for auto-refresh. On connect, after the JWT handshake, it sends a full snapshot (`{"type": "snapshot", "data": {subsystem: state|null, …}}`). On every structural change it broadcasts a per-subsystem delta (`{"type": "versions", "subsystem": …, "data": …}`); on volatile-only changes it sends `{"type": "tick", "subsystem": …, "data": …}`. The client's `handleMessage` patches the matching reactive model in place via `modelSet()`, and the VDOM diff touches only the changed nodes. HTTP remains the fallback for the initial load (3s timer) and for reconnect recovery.
## Zone Model