feat(fleet): GET /api/fleet/capacity returns the scheduler's own survey

Pulled forward from the observability phase because the capacity harness scenario needs it. A test that recomputed the slot arithmetic in bash would drift from `vm_placement` and then agree with itself while the scheduler did something else — the same shape as every silent-success bug in this codebase. Returns `survey()` + `rank()` unmodified, and keeps `unfit` as its own list: "the fleet is full" and "we could not read the fleet" send an operator to different places, so they must not be summed into one number. Co-Authored-By: Claude Opus 5 <[email protected]>
2026-08-09 05:11:21 -07:00
parent d48bdbc9a7
commit dc0443de34
2 changed files with 54 additions and 0 deletions
@@ -402,3 +402,56 @@ async fn bridge_terminal(hub: Arc<NodeHub>, node_id: NodeId, socket: WebSocket)
    }
    hub.terminal_close(node_id, sid).await;
 }
+
+/// `GET /api/fleet/capacity` — what the SCHEDULER sees, verbatim.
+///
+/// Pulled forward from the observability phase because the capacity harness
+/// scenario needs it: a test that recomputed the slot arithmetic in bash would
+/// drift from `vm_placement` and then agree with itself while the scheduler did
+/// something else. This returns `vm_placement::survey` unmodified, so the fleet
+/// page, the harness and the placer cannot disagree.
+///
+/// `backend` narrows to the nodes that can boot one image (`?backend=claude`),
+/// matching what `choose` does for a phase.
+pub async fn capacity(
+    State(state): State<AppState>,
+    Authed(user): Authed,
+    Query(q): Query<CapacityQuery>,
+) -> Result<Json<Value>, ApiError> {
+    let ws = user.workspace_id.as_uuid().to_owned();
+    let (fit, unfit) =
+        crate::vm_placement::survey(&state.pool, &state.node_hub, ws, q.backend.as_deref())
+            .await
+            .map_err(|e| {
+                eprintln!("fleet capacity survey failed: {e}");
+                ApiError::Internal
+            })?;
+    let ranked = crate::vm_placement::rank(fit);
+    Ok(Json(json!({
+        // Total free slots across the fleet. A burst larger than this MUST
+        // queue rather than overcommit — that is the whole feature.
+        "slots": ranked.iter().map(|n| n.slots).sum::<i64>(),
+        "nodes": ranked.iter().map(|n| json!({
+            "id": n.node_id,
+            "name": n.name,
+            "slots": n.slots,
+            "committedVms": n.committed_vms,
+            "headroom": n.headroom,
+            "memTotalMib": n.mem_total_mib,
+            "usedEffMib": n.used_eff_mib,
+            "diskFreeGib": n.disk_free_gib,
+        })).collect::<Vec<_>>(),
+        // Never folded into the above. "Full" and "unreadable" send an
+        // operator to different places, so they stay separate here too.
+        "unfit": unfit.iter().map(|(id, name, why)| json!({
+            "id": id,
+            "name": name,
+            "reason": why.reason(),
+        })).collect::<Vec<_>>(),
+    })))
+}
+
+#[derive(Deserialize)]
+pub struct CapacityQuery {
+    pub backend: Option<String>,
+}