web: poll fleet every 30 s (was 4 s) — SSH load /7.5

specs/tech-architecture/web-dashboard.md

@@ -38,7 +38,7 @@ sm121 web
     ├─ /            embedded UI (include_str! of src/web/ui.html)
     ├─ /api/*       REST (actions + snapshots + raw content)
     ├─ /events      single multiplexed SSE stream
-    ├─ supervisor   tokio task: polls cluster_state::probe every 4 s →
+    ├─ supervisor   tokio task: polls cluster_state::probe every 30 s →
     │               broadcasts state snapshot on the bus
     └─ jobs         serve/stop/power run as tracked tokio tasks that call
                     the existing docker::serve / stop_cluster / power fns
@@ -53,8 +53,9 @@ Key properties:
   `State | Job | Log`. The single `/events` SSE endpoint forwards every
   event; the UI routes by `type`. No per-topic server state.
 - **Supervisor is read-only.** `cluster_state::probe` (what `health` does:
-  `docker ps` per node over SSH) every 4 s. Same SSH load as a manual
-  `health` loop; fine for ≤ a few nodes.
+  `docker ps` per node over SSH) every 30 s. Same SSH load as a manual
+  `health` loop (less); fine for ≤ a few nodes. Job-driven state changes
+  (P2) push immediately, so the 30 s only bounds *external* changes.
 - **Logs.** A log *attach* = one `docker logs -f` SSH session (reusing
   `ssh_session_follow`) whose output is newline-split and pushed to the bus
   as `Log` events, plus a per-attach ring buffer (2000 lines) so new
@@ -95,7 +96,7 @@ Newline-delimited JSON, one event per `data:`:
 {"type":"log","attach":"l2","node":"dgx1.lan","recipe":"qwen38-27b-nvfp4-unsloth","rank":0,"line":"…"}
 ```
 
-- State events: every supervisor tick (4 s) and immediately after any job
+- State events: every supervisor tick (30 s) and immediately after any job
   state change.
 - Reconnect: browser `EventSource` auto-reconnects; on open the client
   re-fetches `GET /api/state` (SSE is fire-and-forget, not a log).

src/web/mod.rs

@@ -2,7 +2,7 @@
 //!
 //! P1 (read-only): `/` serves the embedded UI, `/api/state` the snapshot,
 //! `/events` a multiplexed SSE stream fed by a supervisor task that polls
-//! `cluster_state::probe` every 4 s. Data endpoints are token-gated when
+//! `cluster_state::probe` every 30 s. Data endpoints are token-gated when
 //! `--token` is set (required when binding off-loopback).
 //!
 //! Design: `specs/tech-architecture/web-dashboard.md`.
@@ -30,7 +30,7 @@ use tokio::sync::broadcast;
 use tokio::time::Duration;
 
 const UI_HTML: &str = include_str!("ui.html");
-const POLL_SECS: u64 = 4;
+const POLL_SECS: u64 = 30;
 const BUS_CAP: usize = 1024;
 
 #[derive(Clone)]