fix(insights): make per-category status cards read-only so navigation never blocks on the LLM

Opening an Insights category page froze the whole UI: the per-metric
status card fired a cold, uncapped LLM fetch on mount with no client
abort, so the request hung up to the server's 60s ceiling and the
pending/resolve re-renders thrashed the tab strip and ate touch events.
Timeouts were never negatively cached, so every visit re-blocked.

Fix the class holistically:

- Bound every status GET on the client. The shared use-insight-status
  hook and the inline medication-compliance query now wrap fetch in an
  8s AbortController and stop retrying inline; navigation can no longer
  await an uncapped round-trip.

- Make the status routes read-only. On a cache miss the generator
  enqueues an out-of-band pg-boss job (insight-status-generate) and
  returns a preparing envelope; the client polls a bounded interval
  until the worker warms the cache. Models the dashboard snapshot's
  ready/preparing pattern. The queue is registered in the worker
  allQueues loop + a boss.work handler, with a guard test.

- Fix the nightly warm-pass regression: warm every supported locale
  (the client reads against the active UI locale, not the persisted
  User.locale) and warm on a cached comprehensive outcome too (its
  write still evicts the per-status caches). Add a short-TTL negative
  cache on the timeout path so a transient provider stall no longer
  makes every subsequent visit re-block or re-enqueue.

- Add loading.tsx for the insights segment + the root so any future
  RSC await streams a skeleton instead of blocking the transition.

- Stabilise the advisor regenerate callback so a status-query flip
  can't re-render the memoised tab strip mid-gesture.

New i18n key insights.assessmentPreparing across all six locales.

messages/de.json

@@ -1382,6 +1382,7 @@
     "bmiEmptyDescription": "Für den BMI wird deine Körpergröße benötigt. Hinterlege sie unter Einstellungen → Account.",
     "bmiEmptyAction": "Account-Einstellungen öffnen",
     "noAnalysisYet": "Noch keine Analyse vorhanden.",
+    "assessmentPreparing": "Einschätzung wird vorbereitet…",
     "noProviderConfigured": "Auswertung nicht verfügbar.",
     "startAnalysis": "Analyse starten",
     "refreshAnalysis": "Analyse aktualisieren",

messages/en.json

@@ -1376,6 +1376,7 @@
   },
   "insights": {
     "noAnalysisYet": "No analysis yet.",
+    "assessmentPreparing": "Preparing your assessment…",
     "emptyTitle": "Not enough data yet",
     "emptyDescription": "Insights light up once you've logged a handful of measurements. Start with weight or blood pressure.",
     "emptyAddMeasurement": "Add a measurement",

messages/es.json

@@ -1376,6 +1376,7 @@
   },
   "insights": {
     "noAnalysisYet": "Aún no hay análisis.",
+    "assessmentPreparing": "Preparando tu evaluación…",
     "emptyTitle": "Aún no hay datos suficientes",
     "emptyDescription": "Los insights aparecen en cuanto registres algunas mediciones. Empieza por peso o tensión arterial.",
     "emptyAddMeasurement": "Registrar medición",

messages/fr.json

@@ -1376,6 +1376,7 @@
   },
   "insights": {
     "noAnalysisYet": "Pas encore d’analyse.",
+    "assessmentPreparing": "Préparation de votre évaluation…",
     "emptyTitle": "Pas encore assez de données",
     "emptyDescription": "Les insights apparaissent dès que tu auras saisi quelques mesures. Commence par le poids ou la tension.",
     "emptyAddMeasurement": "Saisir une mesure",

messages/it.json

@@ -1376,6 +1376,7 @@
   },
   "insights": {
     "noAnalysisYet": "Ancora nessuna analisi.",
+    "assessmentPreparing": "Preparazione della valutazione…",
     "emptyTitle": "Dati ancora insufficienti",
     "emptyDescription": "Gli insights compaiono appena registri qualche misurazione. Inizia da peso o pressione.",
     "emptyAddMeasurement": "Registra misurazione",

messages/pl.json

@@ -1376,6 +1376,7 @@
   },
   "insights": {
     "noAnalysisYet": "Brak analizy.",
+    "assessmentPreparing": "Przygotowywanie oceny…",
     "emptyTitle": "Jeszcze za mało danych",
     "emptyDescription": "Insights pojawią się, gdy zarejestrujesz kilka pomiarów. Zacznij od wagi lub ciśnienia.",
     "emptyAddMeasurement": "Zarejestruj pomiar",

src/app/api/insights/blood-pressure-status/route.ts

@@ -23,9 +23,13 @@ export const GET = apiHandler(async (request: NextRequest) => {
   });
   const locale = resolveBloodPressureStatusLocale(resolved);
 
+  // v1.8.3 — read-only: serve the cache, enqueue generation out of band on
+  // a miss. The GET never awaits the provider, so opening /insights/<metric>
+  // can no longer pin the main thread behind a cold LLM round-trip.
   const result = await generateBloodPressureStatusForUser(user.id, {
     locale,
     force: false,
+    readOnly: true,
   });
 
   annotate({ action: { name: "insights.blood-pressure-status" } });

src/app/api/insights/bmi-status/route.ts

@@ -23,9 +23,13 @@ export const GET = apiHandler(async (request: NextRequest) => {
   });
   const locale = resolveBmiStatusLocale(resolved);
 
+  // v1.8.3 — read-only: serve the cache, enqueue generation out of band on
+  // a miss. The GET never awaits the provider, so opening /insights/<metric>
+  // can no longer pin the main thread behind a cold LLM round-trip.
   const result = await generateBmiStatusForUser(user.id, {
     locale,
     force: false,
+    readOnly: true,
   });
 
   annotate({ action: { name: "insights.bmi-status" } });

src/app/api/insights/medication-compliance-status/route.ts

@@ -23,9 +23,13 @@ export const GET = apiHandler(async (request: NextRequest) => {
   });
   const locale = resolveMedicationComplianceStatusLocale(resolved);
 
+  // v1.8.3 — read-only: serve the cache, enqueue generation out of band on
+  // a miss. The GET never awaits the provider, so opening /insights/<metric>
+  // can no longer pin the main thread behind a cold LLM round-trip.
   const result = await generateMedicationComplianceStatusForUser(user.id, {
     locale,
     force: false,
+    readOnly: true,
   });
 
   annotate({ action: { name: "insights.medication-compliance-status" } });

src/app/api/insights/mood-status/route.ts

@@ -23,9 +23,13 @@ export const GET = apiHandler(async (request: NextRequest) => {
   });
   const locale = resolveMoodStatusLocale(resolved);
 
+  // v1.8.3 — read-only: serve the cache, enqueue generation out of band on
+  // a miss. The GET never awaits the provider, so opening /insights/<metric>
+  // can no longer pin the main thread behind a cold LLM round-trip.
   const result = await generateMoodStatusForUser(user.id, {
     locale,
     force: false,
+    readOnly: true,
   });
 
   annotate({ action: { name: "insights.mood-status" } });

src/app/api/insights/pulse-status/route.ts

@@ -23,9 +23,13 @@ export const GET = apiHandler(async (request: NextRequest) => {
   });
   const locale = resolvePulseStatusLocale(resolved);
 
+  // v1.8.3 — read-only: serve the cache, enqueue generation out of band on
+  // a miss. The GET never awaits the provider, so opening /insights/<metric>
+  // can no longer pin the main thread behind a cold LLM round-trip.
   const result = await generatePulseStatusForUser(user.id, {
     locale,
     force: false,
+    readOnly: true,
   });
 
   annotate({ action: { name: "insights.pulse-status" } });

src/app/api/insights/weight-status/route.ts

@@ -23,9 +23,13 @@ export const GET = apiHandler(async (request: NextRequest) => {
   });
   const locale = resolveWeightStatusLocale(resolved);
 
+  // v1.8.3 — read-only: serve the cache, enqueue generation out of band on
+  // a miss. The GET never awaits the provider, so opening /insights/<metric>
+  // can no longer pin the main thread behind a cold LLM round-trip.
   const result = await generateWeightStatusForUser(user.id, {
     locale,
     force: false,
+    readOnly: true,
   });
 
   annotate({ action: { name: "insights.weight-status" } });

src/app/insights/blood-pressure/page.tsx

@@ -124,6 +124,7 @@ export default function InsightsBlutdruckPage() {
         cached={status?.cached ?? false}
         updatedAt={status?.updatedAt ?? null}
         loading={isStatusLoading}
+        preparing={status?.preparing ?? false}
       />
 
       <CoachLaunchButton />

src/app/insights/bmi/page.tsx

@@ -133,6 +133,7 @@ export default function InsightsBmiPage() {
         cached={status?.cached ?? false}
         updatedAt={status?.updatedAt ?? null}
         loading={isStatusLoading}
+        preparing={status?.preparing ?? false}
       />
 
       <CoachLaunchButton />

src/app/insights/loading.tsx

@@ -0,0 +1,37 @@
+/**
+ * v1.8.3 — route-level loading skeleton for every `/insights/*` segment.
+ *
+ * Defense-in-depth for the navigation-freeze class: any RSC await on an
+ * insights route (now or in a future refactor) streams this skeleton
+ * instead of blocking the transition. The geometry mirrors `SubPageShell`
+ * (heading + chart card + assessment card) so the post-load swap reflows
+ * by less than a row. Pure server component — no client hooks, no data
+ * fetch; Next.js renders it instantly while the segment streams.
+ */
+export default function InsightsLoading() {
+  return (
+    <div data-slot="insights-subpage-loading" className="space-y-6">
+      <header className="space-y-2">
+        <div className="bg-muted h-7 w-48 animate-pulse rounded motion-reduce:animate-none" />
+        <div className="bg-muted/70 h-4 w-72 animate-pulse rounded motion-reduce:animate-none" />
+      </header>
+
+      {/* Chart card */}
+      <div className="bg-card rounded-xl border p-4">
+        <div className="bg-muted mb-4 h-4 w-32 animate-pulse rounded motion-reduce:animate-none" />
+        <div className="bg-muted/60 h-56 w-full animate-pulse rounded-lg motion-reduce:animate-none" />
+      </div>
+
+      {/* Assessment card */}
+      <div className="bg-card space-y-2 rounded-xl border p-4">
+        <div className="flex items-center gap-2">
+          <div className="bg-muted h-5 w-5 animate-pulse rounded motion-reduce:animate-none" />
+          <div className="bg-muted h-4 w-40 animate-pulse rounded motion-reduce:animate-none" />
+        </div>
+        <div className="bg-muted h-3.5 w-full animate-pulse rounded motion-reduce:animate-none" />
+        <div className="bg-muted h-3.5 w-11/12 animate-pulse rounded motion-reduce:animate-none" />
+        <div className="bg-muted h-3.5 w-9/12 animate-pulse rounded motion-reduce:animate-none" />
+      </div>
+    </div>
+  );
+}

src/app/insights/medications/page.tsx

@@ -58,8 +58,18 @@ interface MedicationComplianceStatusData {
   medications: Array<{ medicationId: string; text: string }>;
   cached: boolean;
   updatedAt: string | null;
+  // v1.8.3 — read-only route returns preparing:true with summary:null on a
+  // cache miss while the worker warms the assessment; the card polls until
+  // it lands. See `use-insight-status.ts` for the shared rationale.
+  preparing?: boolean;
 }
 
+// v1.8.3 — client ceiling + preparing-poll cadence, mirroring the shared
+// `use-insight-status` hook so the medication-compliance card never awaits
+// an uncapped LLM round-trip on navigation.
+const MED_STATUS_TIMEOUT_MS = 8_000;
+const MED_STATUS_POLL_MS = 4_000;
+
 interface MedicationDailyData {
   expected: number;
   taken: number;
@@ -91,15 +101,28 @@ export default function InsightsMedikamentePage() {
   const { data: status, isLoading: isStatusLoading } = useQuery({
     queryKey: queryKeys.insightsMedicationComplianceStatus(locale),
     queryFn: async () => {
-      const res = await fetch(
-        `/api/insights/medication-compliance-status?locale=${locale}`,
+      const controller = new AbortController();
+      const timeoutHandle = setTimeout(
+        () => controller.abort(),
+        MED_STATUS_TIMEOUT_MS,
       );
-      if (!res.ok) throw new Error("Failed");
-      const json = await res.json();
-      return json.data as MedicationComplianceStatusData;
+      try {
+        const res = await fetch(
+          `/api/insights/medication-compliance-status?locale=${locale}`,
+          { signal: controller.signal },
+        );
+        if (!res.ok) throw new Error("Failed");
+        const json = await res.json();
+        return json.data as MedicationComplianceStatusData;
+      } finally {
+        clearTimeout(timeoutHandle);
+      }
     },
     enabled: isAuthenticated,
     staleTime: 60 * 1000,
+    retry: 0,
+    refetchInterval: (query) =>
+      query.state.data?.preparing ? MED_STATUS_POLL_MS : false,
   });
 
   const medications = comprehensive?.medications ?? [];
@@ -253,6 +276,7 @@ export default function InsightsMedikamentePage() {
         cached={status?.cached ?? false}
         updatedAt={status?.updatedAt ?? null}
         loading={isStatusLoading}
+        preparing={status?.preparing ?? false}
       />
 
       {/* v1.4.25 W4d — GLP-1 therapy timeline. Self-hides for users

src/app/insights/mood/page.tsx

@@ -115,6 +115,7 @@ export default function InsightsStimmungPage() {
         cached={status?.cached ?? false}
         updatedAt={status?.updatedAt ?? null}
         loading={isStatusLoading}
+        preparing={status?.preparing ?? false}
       />
 
       <CoachLaunchButton />

src/app/insights/pulse/page.tsx

@@ -136,6 +136,7 @@ export default function InsightsPulsPage() {
         cached={status?.cached ?? false}
         updatedAt={status?.updatedAt ?? null}
         loading={isStatusLoading}
+        preparing={status?.preparing ?? false}
       />
 
       {/* v1.4.25 W16a — VO2 max sits on the cardio sub-page because it

src/app/insights/weight/page.tsx

@@ -97,6 +97,7 @@ export default function InsightsGewichtPage() {
         cached={status?.cached ?? false}
         updatedAt={status?.updatedAt ?? null}
         loading={isStatusLoading}
+        preparing={status?.preparing ?? false}
       />
 
       <CoachLaunchButton />

src/app/loading.tsx

@@ -0,0 +1,24 @@
+/**
+ * v1.8.3 — root-level loading skeleton (dashboard segment + any top-level
+ * route without its own `loading.tsx`).
+ *
+ * Defense-in-depth for the navigation-freeze class: a segment that streams
+ * paints this tile-grid skeleton instead of a blank frame. Pure server
+ * component — no client hooks, no data fetch. More specific segments (e.g.
+ * `/insights`) provide their own skeleton and take precedence.
+ */
+export default function RootLoading() {
+  return (
+    <div data-slot="dashboard-loading" className="space-y-6 p-4">
+      <div className="bg-muted h-7 w-44 animate-pulse rounded motion-reduce:animate-none" />
+      <div className="grid grid-cols-1 gap-4 sm:grid-cols-2 lg:grid-cols-3">
+        {Array.from({ length: 6 }).map((_, i) => (
+          <div key={i} className="bg-card rounded-xl border p-4">
+            <div className="bg-muted mb-3 h-4 w-24 animate-pulse rounded motion-reduce:animate-none" />
+            <div className="bg-muted/60 h-24 w-full animate-pulse rounded-lg motion-reduce:animate-none" />
+          </div>
+        ))}
+      </div>
+    </div>
+  );
+}

src/components/insights/insight-status-card.tsx

@@ -18,6 +18,13 @@ interface InsightStatusCardProps {
   cached: boolean;
   updatedAt: string | null;
   loading?: boolean;
+  /**
+   * v1.8.3 — the read-only status route enqueued an out-of-band generation
+   * and the assessment isn't warm yet. Render the same skeleton geometry as
+   * `loading` but with a "preparing" caption so the user understands the
+   * card is being assembled, not stuck. The client polls until text lands.
+   */
+  preparing?: boolean;
 }
 
 // ─── Main Component ───────────────────────────────────────
@@ -30,6 +37,7 @@ export function InsightStatusCard({
   cached,
   updatedAt,
   loading = false,
+  preparing = false,
 }: InsightStatusCardProps) {
   const { t } = useTranslations();
   const flags = useFeatureFlags();
@@ -38,6 +46,36 @@ export function InsightStatusCard({
   // suppressed in full so the layout collapses naturally.
   if (!flags.insightStatus) return null;
 
+  // v1.8.3 — preparing: the read-only route enqueued a generation and the
+  // client is polling. Render the loading skeleton geometry with a
+  // preparing caption so the card reads as "assembling", not stuck. Only
+  // shown when a provider is configured (the route returns the no-key
+  // fallback otherwise).
+  if (preparing && hasProvider && !text) {
+    return (
+      <Card
+        aria-busy="true"
+        aria-live="polite"
+        data-testid="insight-status-card-preparing"
+      >
+        <CardHeader className="pb-2">
+          <div className="flex items-center gap-2">
+            {icon}
+            <CardTitle className="text-base">{title}</CardTitle>
+          </div>
+        </CardHeader>
+        <CardContent className="space-y-2">
+          <div className="bg-muted h-3.5 w-full animate-pulse rounded motion-reduce:animate-none" />
+          <div className="bg-muted h-3.5 w-11/12 animate-pulse rounded motion-reduce:animate-none" />
+          <div className="bg-muted h-3.5 w-9/12 animate-pulse rounded motion-reduce:animate-none" />
+          <p className="text-muted-foreground text-xs">
+            {t("insights.assessmentPreparing")}
+          </p>
+        </CardContent>
+      </Card>
+    );
+  }
+
   if (loading) {
     // v1.4.37 — structured skeleton over the centred spinner. When the
     // per-metric status route had to fall back to the 20 s provider

src/components/insights/use-insights-advisor.ts

@@ -1,5 +1,6 @@
 "use client";
 
+import { useCallback } from "react";
 import { useMutation, useQuery, useQueryClient } from "@tanstack/react-query";
 import type { InsightResult } from "@/lib/ai/types";
 import {
@@ -176,12 +177,20 @@ export function useInsightsAdvisorQuery(
     },
   });
 
+  // v1.8.3 — stabilise the `regenerate` callback so the memoised
+  // `<InsightsTabStrip>` (which receives it as `onRegenerate`) is not
+  // re-rendered on every shell render by a fresh arrow reference. A
+  // status-query flip on a sub-page that re-renders the shell would
+  // otherwise re-reconcile the whole strip mid-gesture and eat taps.
+  const { mutate } = mutation;
+  const regenerate = useCallback(() => mutate(), [mutate]);
+
   return {
     payload: query.data ?? null,
     isLoading: query.isLoading,
     isError: query.isError,
     error: (query.error as Error | null) ?? null,
-    regenerate: () => mutation.mutate(),
+    regenerate,
     isRegenerating: mutation.isPending,
     regenerateError: (mutation.error as Error | null) ?? null,
   };

src/hooks/use-insight-status.ts

@@ -23,8 +23,32 @@ export interface InsightStatusData {
   text: string | null;
   cached: boolean;
   updatedAt: string | null;
+  /**
+   * v1.8.3 — the route is read-only: a cache miss enqueues an out-of-band
+   * generation and returns `preparing: true` with `text: null`. The card
+   * shows a preparing state and the hook polls (bounded) until the worker
+   * warms the cache. Absent / false means the payload is terminal.
+   */
+  preparing?: boolean;
 }
 
+/**
+ * v1.8.3 — client-side ceiling on the status GET. The route is now
+ * read-only (sub-second cache read), so a slow response means the network
+ * itself is degraded, not the LLM. Mirror the advisor hook's 8 s
+ * `ADVISOR_TIMEOUT_MS`: abort and surface the empty / preparing state
+ * rather than letting a hung request pin the navigation. No user
+ * navigation may ever await an uncapped round-trip.
+ */
+const STATUS_TIMEOUT_MS = 8_000;
+
+/**
+ * Poll interval while a card is `preparing`. Bounded by `staleTime` so a
+ * settled (terminal) payload never re-fetches on a timer; only the
+ * preparing state keeps polling, and it stops as soon as text lands.
+ */
+const STATUS_POLL_MS = 4_000;
+
 /**
  * Metric slugs the sub-pages render. Each slug is paired with the
  * Insights-status endpoint and the matching `queryKeys.*` factory so
@@ -71,17 +95,37 @@ export function useInsightStatus(metric: InsightStatusMetric) {
   return useQuery({
     queryKey: QUERY_KEY_FACTORY[metric](locale),
     queryFn: async (): Promise<InsightStatusData> => {
-      const res = await fetch(`/api/insights/${metric}-status?locale=${locale}`);
-      if (!res.ok) throw new Error("Failed");
-      const json = (await res.json()) as { data: InsightStatusData };
-      return json.data;
+      // v1.8.3 — bound the fetch on the client. AbortController + 8 s
+      // timeout so a degraded network can't pin the navigation thread.
+      // The card surfaces its preparing / empty state on abort exactly as
+      // it does for a `preparing` payload, and the bounded poll retries.
+      const controller = new AbortController();
+      const timeoutHandle = setTimeout(
+        () => controller.abort(),
+        STATUS_TIMEOUT_MS,
+      );
+      try {
+        const res = await fetch(
+          `/api/insights/${metric}-status?locale=${locale}`,
+          { signal: controller.signal },
+        );
+        if (!res.ok) throw new Error("Failed");
+        const json = (await res.json()) as { data: InsightStatusData };
+        return json.data;
+      } finally {
+        clearTimeout(timeoutHandle);
+      }
     },
     enabled: isAuthenticated,
     staleTime: 60 * 1000,
-    // v1.4.28 FB-D2 — the status route now returns a deterministic
-    // cached-style envelope on provider timeout. Retrying that response
-    // re-fires the same expensive upstream call and the user perceives
-    // a longer hang. Drop retries to surface the fallback immediately.
+    // v1.4.28 FB-D2 — the status route returns a deterministic envelope on
+    // a miss/timeout. Retrying inline re-fires work and lengthens the
+    // perceived hang; the bounded `preparing` poll below covers the warm-up.
     retry: 0,
+    // v1.8.3 — poll only while the worker is preparing the assessment.
+    // Returns false (no timer) for any terminal payload so a settled card
+    // never re-fetches on an interval.
+    refetchInterval: (query) =>
+      query.state.data?.preparing ? STATUS_POLL_MS : false,
   });
 }

src/lib/insights/__tests__/blood-pressure-status.test.ts

@@ -152,7 +152,19 @@ describe("generateBloodPressureStatusForUser — timeout/error never persists",
     expect(result.text).toBeTruthy();
     expect(result.cached).toBe(true);
     expect(result.updatedAt).toBeNull();
-    expect(prisma.auditLog.create).not.toHaveBeenCalled();
+    // v1.8.3 — no real assessment persisted (updatedAt stays null above),
+    // but a short-TTL negative stub IS written so the read-only route does
+    // not re-enqueue on every navigation while the provider is degraded.
+    // The stub is a timeout marker that `readFreshStatusText` rejects.
+    await Promise.resolve();
+    for (const call of vi.mocked(prisma.auditLog.create).mock.calls) {
+      const details = JSON.parse(
+        (call[0] as { data: { details: string } }).data.details,
+      );
+      expect(details.timeout === true || details.model === "timeout-stub").toBe(
+        true,
+      );
+    }
   });
 });
 

src/lib/insights/__tests__/bmi-status.test.ts

@@ -121,7 +121,19 @@ describe("generateBmiStatusForUser — timeout/error never persists", () => {
     expect(result.text).toBeTruthy();
     expect(result.cached).toBe(true);
     expect(result.updatedAt).toBeNull();
-    expect(prisma.auditLog.create).not.toHaveBeenCalled();
+    // v1.8.3 — no real assessment persisted (updatedAt stays null above),
+    // but a short-TTL negative stub IS written so the read-only route does
+    // not re-enqueue on every navigation while the provider is degraded.
+    // The stub is a timeout marker that `readFreshStatusText` rejects.
+    await Promise.resolve();
+    for (const call of vi.mocked(prisma.auditLog.create).mock.calls) {
+      const details = JSON.parse(
+        (call[0] as { data: { details: string } }).data.details,
+      );
+      expect(details.timeout === true || details.model === "timeout-stub").toBe(
+        true,
+      );
+    }
   });
 
   it("serves the fallback without writing a cache row on provider error", async () => {
@@ -140,7 +152,19 @@ describe("generateBmiStatusForUser — timeout/error never persists", () => {
     const result = await generateBmiStatusForUser("user-1", { locale: "en" });
     expect(result.text).toBeTruthy();
     expect(result.updatedAt).toBeNull();
-    expect(prisma.auditLog.create).not.toHaveBeenCalled();
+    // v1.8.3 — no real assessment persisted (updatedAt stays null above),
+    // but a short-TTL negative stub IS written so the read-only route does
+    // not re-enqueue on every navigation while the provider is degraded.
+    // The stub is a timeout marker that `readFreshStatusText` rejects.
+    await Promise.resolve();
+    for (const call of vi.mocked(prisma.auditLog.create).mock.calls) {
+      const details = JSON.parse(
+        (call[0] as { data: { details: string } }).data.details,
+      );
+      expect(details.timeout === true || details.model === "timeout-stub").toBe(
+        true,
+      );
+    }
   });
 });
 

src/lib/insights/__tests__/general-status.test.ts

@@ -162,7 +162,19 @@ describe("generateGeneralStatusForUser — timeout/error never persists", () =>
     expect(result.text).toBeTruthy();
     expect(result.cached).toBe(true);
     expect(result.updatedAt).toBeNull();
-    expect(prisma.auditLog.create).not.toHaveBeenCalled();
+    // v1.8.3 — no real assessment persisted (updatedAt stays null above),
+    // but a short-TTL negative stub IS written so the read-only route does
+    // not re-enqueue on every navigation while the provider is degraded.
+    // The stub is a timeout marker that `readFreshStatusText` rejects.
+    await Promise.resolve();
+    for (const call of vi.mocked(prisma.auditLog.create).mock.calls) {
+      const details = JSON.parse(
+        (call[0] as { data: { details: string } }).data.details,
+      );
+      expect(details.timeout === true || details.model === "timeout-stub").toBe(
+        true,
+      );
+    }
   });
 });
 

src/lib/insights/__tests__/medication-compliance-status.test.ts

@@ -155,7 +155,19 @@ describe("generateMedicationComplianceStatusForUser — timeout never persists",
     expect(result.medications).toEqual([]);
     expect(result.cached).toBe(true);
     expect(result.updatedAt).toBeNull();
-    expect(prisma.auditLog.create).not.toHaveBeenCalled();
+    // v1.8.3 — no real assessment persisted (updatedAt stays null above),
+    // but a short-TTL negative stub IS written so the read-only route does
+    // not re-enqueue on every navigation while the provider is degraded.
+    // The stub is a timeout marker that `readFreshStatusText` rejects.
+    await Promise.resolve();
+    for (const call of vi.mocked(prisma.auditLog.create).mock.calls) {
+      const details = JSON.parse(
+        (call[0] as { data: { details: string } }).data.details,
+      );
+      expect(details.timeout === true || details.model === "timeout-stub").toBe(
+        true,
+      );
+    }
   });
 });
 

src/lib/insights/__tests__/mood-status.test.ts

@@ -118,7 +118,19 @@ describe("generateMoodStatusForUser — timeout/error never persists", () => {
     expect(result.text).toBeTruthy();
     expect(result.cached).toBe(true);
     expect(result.updatedAt).toBeNull();
-    expect(prisma.auditLog.create).not.toHaveBeenCalled();
+    // v1.8.3 — no real assessment persisted (updatedAt stays null above),
+    // but a short-TTL negative stub IS written so the read-only route does
+    // not re-enqueue on every navigation while the provider is degraded.
+    // The stub is a timeout marker that `readFreshStatusText` rejects.
+    await Promise.resolve();
+    for (const call of vi.mocked(prisma.auditLog.create).mock.calls) {
+      const details = JSON.parse(
+        (call[0] as { data: { details: string } }).data.details,
+      );
+      expect(details.timeout === true || details.model === "timeout-stub").toBe(
+        true,
+      );
+    }
   });
 });
 

src/lib/insights/__tests__/pulse-status-timeout.test.ts

@@ -4,12 +4,15 @@ import { describe, expect, it, vi, beforeEach } from "vitest";
  * Provider-timeout fallback for the pulse-status route. When the
  * provider chain does not resolve inside `STATUS_PROVIDER_TIMEOUT_MS`
  * the route returns the deterministic no-key fallback so the
- * InsightStatusCard renders instead of spinning — but it must NOT
- * persist the fallback. The earlier behaviour cached a `timeout-stub`
- * row keyed to today; that stub stuck until midnight and hid the real
- * data-driven assessment. The fix treats a timeout as a transient miss:
- * serve the fallback for this render, write nothing, re-attempt next
- * mount.
+ * InsightStatusCard renders instead of spinning. It must NOT persist the
+ * fallback AS AN ASSESSMENT (the pre-v1.4.28 stick-until-midnight bug);
+ * `updatedAt` stays null and the served text is never a real assessment.
+ *
+ * v1.8.3 — the timeout path now writes a *short-TTL negative stub*
+ * (`{ timeout:true, model:"timeout-stub", retryAt }`). It is explicitly
+ * rejected by `readFreshStatusText`, so it can never hide the real
+ * assessment; its sole purpose is to stop the read-only route re-enqueuing
+ * generation on every navigation while a provider is degraded.
  */
 
 vi.mock("@/lib/db", () => ({
@@ -64,8 +67,23 @@ describe("generatePulseStatusForUser — provider timeout fallback", () => {
     expect(result.cached).toBe(true);
     expect(typeof result.text).toBe("string");
     expect(result.text?.length ?? 0).toBeGreaterThan(0);
-    // No persisted row — the sticky-stub bug is gone.
+    // No real assessment persisted — `updatedAt` stays null so the card
+    // never mislabels the fallback as a fresh assessment.
     expect(result.updatedAt).toBeNull();
-    expect(prisma.auditLog.create).not.toHaveBeenCalled();
+
+    // v1.8.3 — a short-TTL negative stub IS persisted (fire-and-forget) so
+    // the read-only route doesn't re-enqueue on every navigation while the
+    // provider is degraded. It is marked as a timeout stub, which
+    // `readFreshStatusText` rejects, so it never hides the real assessment.
+    // The write is fire-and-forget (`void`), so flush the microtask queue.
+    await Promise.resolve();
+    expect(prisma.auditLog.create).toHaveBeenCalledTimes(1);
+    const persisted = vi.mocked(prisma.auditLog.create).mock.calls[0][0] as {
+      data: { details: string };
+    };
+    const stub = JSON.parse(persisted.data.details);
+    expect(stub.timeout).toBe(true);
+    expect(stub.model).toBe("timeout-stub");
+    expect(typeof stub.retryAt).toBe("string");
   });
 });

src/lib/insights/__tests__/status-cache.test.ts

@@ -6,8 +6,22 @@ vi.mock("@/lib/db", () => ({
   },
 }));
 
+const hasUsableStatusProvider = vi.fn();
+vi.mock("@/lib/insights/status-provider", () => ({
+  hasUsableStatusProvider: (...a: unknown[]) => hasUsableStatusProvider(...a),
+}));
+
+const enqueueStatusGeneration = vi.fn();
+vi.mock("@/lib/jobs/insight-status-generate-shared", () => ({
+  enqueueStatusGeneration: (...a: unknown[]) => enqueueStatusGeneration(...a),
+}));
+
 import { prisma } from "@/lib/db";
-import { isTimeoutStub, readFreshStatusText } from "../status-cache";
+import {
+  isTimeoutStub,
+  readFreshStatusText,
+  resolveReadOnlyStatusMiss,
+} from "../status-cache";
 
 const TODAY = "2026-05-31";
 
@@ -121,3 +135,71 @@ describe("readFreshStatusText", () => {
     expect(hit).toBeNull();
   });
 });
+
+describe("resolveReadOnlyStatusMiss", () => {
+  it("returns no-provider without enqueuing when the user has no provider", async () => {
+    hasUsableStatusProvider.mockResolvedValue(false);
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId: "u1",
+      metric: "weight",
+      locale: "en",
+    });
+    expect(outcome).toBe("no-provider");
+    expect(enqueueStatusGeneration).not.toHaveBeenCalled();
+  });
+
+  it("enqueues generation and returns preparing on a clean miss", async () => {
+    hasUsableStatusProvider.mockResolvedValue(true);
+    // No negative stub present.
+    vi.mocked(prisma.auditLog.findFirst).mockResolvedValue(null as never);
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId: "u1",
+      metric: "pulse",
+      locale: "de",
+    });
+    expect(outcome).toBe("preparing");
+    expect(enqueueStatusGeneration).toHaveBeenCalledWith({
+      userId: "u1",
+      metric: "pulse",
+      locale: "de",
+    });
+  });
+
+  it("suppresses re-enqueue while a fresh timeout stub exists", async () => {
+    hasUsableStatusProvider.mockResolvedValue(true);
+    vi.mocked(prisma.auditLog.findFirst).mockResolvedValue(
+      cacheRow({
+        dateKey: TODAY,
+        timeout: true,
+        model: "timeout-stub",
+        retryAt: new Date(Date.now() + 60_000).toISOString(),
+      }) as never,
+    );
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId: "u1",
+      metric: "mood",
+      locale: "en",
+    });
+    expect(outcome).toBe("preparing");
+    expect(enqueueStatusGeneration).not.toHaveBeenCalled();
+  });
+
+  it("re-enqueues once the negative stub's retryAt has passed", async () => {
+    hasUsableStatusProvider.mockResolvedValue(true);
+    vi.mocked(prisma.auditLog.findFirst).mockResolvedValue(
+      cacheRow({
+        dateKey: TODAY,
+        timeout: true,
+        model: "timeout-stub",
+        retryAt: new Date(Date.now() - 60_000).toISOString(),
+      }) as never,
+    );
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId: "u1",
+      metric: "bmi",
+      locale: "de",
+    });
+    expect(outcome).toBe("preparing");
+    expect(enqueueStatusGeneration).toHaveBeenCalledTimes(1);
+  });
+});

src/lib/insights/__tests__/weight-status.test.ts

@@ -126,7 +126,19 @@ describe("generateWeightStatusForUser — timeout/error never persists", () => {
     expect(result.text).toBeTruthy();
     expect(result.cached).toBe(true);
     expect(result.updatedAt).toBeNull();
-    expect(prisma.auditLog.create).not.toHaveBeenCalled();
+    // v1.8.3 — no real assessment persisted (updatedAt stays null above),
+    // but a short-TTL negative stub IS written so the read-only route does
+    // not re-enqueue on every navigation while the provider is degraded.
+    // The stub is a timeout marker that `readFreshStatusText` rejects.
+    await Promise.resolve();
+    for (const call of vi.mocked(prisma.auditLog.create).mock.calls) {
+      const details = JSON.parse(
+        (call[0] as { data: { details: string } }).data.details,
+      );
+      expect(details.timeout === true || details.model === "timeout-stub").toBe(
+        true,
+      );
+    }
   });
 });
 

src/lib/insights/blood-pressure-status.ts

@@ -42,7 +42,10 @@ import {
   summarizeSeries,
 } from "@/lib/insights/status-shared";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
-import { readFreshStatusText } from "@/lib/insights/status-cache";
+import {
+  readFreshStatusText,
+  resolveReadOnlyStatusMiss,
+} from "@/lib/insights/status-cache";
 import { returnTimeoutFallback } from "@/lib/insights/timeout-fallback";
 import { annotate } from "@/lib/logging/context";
 import { toBerlinDayKey } from "@/lib/tz/resolver";
@@ -94,15 +97,19 @@ export async function generateBloodPressureStatusForUser(
   options?: {
     locale?: string | null;
     force?: boolean;
+    /** v1.8.3 — read-only navigation path; see weight-status for the rationale. */
+    readOnly?: boolean;
   },
 ): Promise<{
   hasProvider: boolean;
   text: string | null;
   cached: boolean;
   updatedAt: string | null;
+  preparing?: boolean;
 }> {
   const locale = normalizeLocale(options?.locale);
   const force = options?.force === true;
+  const readOnly = options?.readOnly === true;
   const cacheAction = `insights.blood-pressure-status.${locale}`;
   const todayKey = toBerlinDayKey(new Date());
 
@@ -121,6 +128,29 @@ export async function generateBloodPressureStatusForUser(
     };
   }
 
+  if (readOnly) {
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId,
+      metric: "blood-pressure",
+      locale,
+    });
+    if (outcome === "no-provider") {
+      return {
+        hasProvider: false,
+        text: getNoKeyBloodPressureStatusText(locale),
+        cached: true,
+        updatedAt: null,
+      };
+    }
+    return {
+      hasProvider: true,
+      text: null,
+      cached: false,
+      updatedAt: null,
+      preparing: true,
+    };
+  }
+
   const user = await prisma.user.findUnique({
     where: { id: userId },
     select: {
@@ -523,6 +553,8 @@ export async function generateBloodPressureStatusForUser(
     return returnTimeoutFallback({
       cacheAction,
       reason: outcome.kind,
+      userId,
+      todayKey,
       stubText: getNoKeyBloodPressureStatusText(locale),
     });
   }

src/lib/insights/bmi-status.ts

@@ -22,7 +22,10 @@ import {
   summarizeSeries,
 } from "@/lib/insights/status-shared";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
-import { readFreshStatusText } from "@/lib/insights/status-cache";
+import {
+  readFreshStatusText,
+  resolveReadOnlyStatusMiss,
+} from "@/lib/insights/status-cache";
 import { returnTimeoutFallback } from "@/lib/insights/timeout-fallback";
 import { annotate } from "@/lib/logging/context";
 import { toBerlinDayKey } from "@/lib/tz/resolver";
@@ -32,15 +35,19 @@ export async function generateBmiStatusForUser(
   options?: {
     locale?: string | null;
     force?: boolean;
+    /** v1.8.3 — read-only navigation path; see weight-status for the rationale. */
+    readOnly?: boolean;
   },
 ): Promise<{
   hasProvider: boolean;
   text: string | null;
   cached: boolean;
   updatedAt: string | null;
+  preparing?: boolean;
 }> {
   const locale = normalizeLocale(options?.locale);
   const force = options?.force === true;
+  const readOnly = options?.readOnly === true;
   const cacheAction = `insights.bmi-status.${locale}`;
   const todayKey = toBerlinDayKey(new Date());
 
@@ -59,6 +66,29 @@ export async function generateBmiStatusForUser(
     };
   }
 
+  if (readOnly) {
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId,
+      metric: "bmi",
+      locale,
+    });
+    if (outcome === "no-provider") {
+      return {
+        hasProvider: false,
+        text: getNoKeyBmiStatusText(locale),
+        cached: true,
+        updatedAt: null,
+      };
+    }
+    return {
+      hasProvider: true,
+      text: null,
+      cached: false,
+      updatedAt: null,
+      preparing: true,
+    };
+  }
+
   const user = await prisma.user.findUnique({
     where: { id: userId },
     select: {
@@ -227,6 +257,8 @@ export async function generateBmiStatusForUser(
     return returnTimeoutFallback({
       cacheAction,
       reason: outcome.kind,
+      userId,
+      todayKey,
       stubText: getNoKeyBmiStatusText(locale),
     });
   }

src/lib/insights/general-status.ts

@@ -25,7 +25,10 @@ import {
   summarizeSeries,
 } from "@/lib/insights/status-shared";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
-import { readFreshStatusText } from "@/lib/insights/status-cache";
+import {
+  readFreshStatusText,
+  resolveReadOnlyStatusMiss,
+} from "@/lib/insights/status-cache";
 import { returnTimeoutFallback } from "@/lib/insights/timeout-fallback";
 import { annotate } from "@/lib/logging/context";
 import { toBerlinDayKey } from "@/lib/tz/resolver";
@@ -47,15 +50,19 @@ export async function generateGeneralStatusForUser(
   options?: {
     locale?: string | null;
     force?: boolean;
+    /** v1.8.3 — read-only navigation path; see weight-status for the rationale. */
+    readOnly?: boolean;
   },
 ): Promise<{
   hasProvider: boolean;
   text: string | null;
   cached: boolean;
   updatedAt: string | null;
+  preparing?: boolean;
 }> {
   const locale = normalizeLocale(options?.locale);
   const force = options?.force === true;
+  const readOnly = options?.readOnly === true;
   const cacheAction = `insights.general-status.${locale}`;
   const todayKey = toBerlinDayKey(new Date());
 
@@ -74,6 +81,29 @@ export async function generateGeneralStatusForUser(
     };
   }
 
+  if (readOnly) {
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId,
+      metric: "general",
+      locale,
+    });
+    if (outcome === "no-provider") {
+      return {
+        hasProvider: false,
+        text: getNoKeyGeneralStatusText(locale),
+        cached: true,
+        updatedAt: null,
+      };
+    }
+    return {
+      hasProvider: true,
+      text: null,
+      cached: false,
+      updatedAt: null,
+      preparing: true,
+    };
+  }
+
   const user = await prisma.user.findUnique({
     where: { id: userId },
     select: {
@@ -348,6 +378,8 @@ export async function generateGeneralStatusForUser(
     return returnTimeoutFallback({
       cacheAction,
       reason: outcome.kind,
+      userId,
+      todayKey,
       stubText: getNoKeyGeneralStatusText(locale),
     });
   }

src/lib/insights/medication-compliance-status.ts

@@ -27,7 +27,10 @@ import {
   round,
 } from "@/lib/insights/status-shared";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
-import { isTimeoutStub } from "@/lib/insights/status-cache";
+import {
+  isTimeoutStub,
+  resolveReadOnlyStatusMiss,
+} from "@/lib/insights/status-cache";
 import { returnTimeoutFallback } from "@/lib/insights/timeout-fallback";
 import { annotate } from "@/lib/logging/context";
 import { toBerlinDayKey } from "@/lib/tz/resolver";
@@ -46,6 +49,8 @@ export async function generateMedicationComplianceStatusForUser(
   options?: {
     locale?: string | null;
     force?: boolean;
+    /** v1.8.3 — read-only navigation path; see weight-status for the rationale. */
+    readOnly?: boolean;
   },
 ): Promise<{
   hasProvider: boolean;
@@ -53,9 +58,11 @@ export async function generateMedicationComplianceStatusForUser(
   medications: MedicationSummaryItem[];
   cached: boolean;
   updatedAt: string | null;
+  preparing?: boolean;
 }> {
   const locale = normalizeLocale(options?.locale);
   const force = options?.force === true;
+  const readOnly = options?.readOnly === true;
   const cacheAction = `insights.medication-compliance-status.${locale}`;
   const todayKey = toBerlinDayKey(new Date());
 
@@ -106,6 +113,33 @@ export async function generateMedicationComplianceStatusForUser(
     }
   }
 
+  // v1.8.3 — read-only navigation path: never block on the provider.
+  // Enqueue generation out of band and return preparing / no-provider.
+  if (readOnly) {
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId,
+      metric: "medication-compliance",
+      locale,
+    });
+    if (outcome === "no-provider") {
+      return {
+        hasProvider: false,
+        summary: getNoKeyMedicationComplianceStatusText(locale),
+        medications: [],
+        cached: true,
+        updatedAt: null,
+      };
+    }
+    return {
+      hasProvider: true,
+      summary: null,
+      medications: [],
+      cached: false,
+      updatedAt: null,
+      preparing: true,
+    };
+  }
+
   const medications = await prisma.medication.findMany({
     where: { userId, active: true },
     include: { schedules: true },
@@ -338,6 +372,8 @@ export async function generateMedicationComplianceStatusForUser(
     returnTimeoutFallback({
       cacheAction,
       reason: outcome.kind,
+      userId,
+      todayKey,
       stubText: getNoKeyMedicationComplianceStatusText(locale),
     });
     return {

src/lib/insights/mood-status.ts

@@ -28,7 +28,10 @@ import {
   summarizeSeries,
 } from "@/lib/insights/status-shared";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
-import { readFreshStatusText } from "@/lib/insights/status-cache";
+import {
+  readFreshStatusText,
+  resolveReadOnlyStatusMiss,
+} from "@/lib/insights/status-cache";
 import { returnTimeoutFallback } from "@/lib/insights/timeout-fallback";
 import { annotate } from "@/lib/logging/context";
 import { toBerlinDayKey } from "@/lib/tz/resolver";
@@ -70,15 +73,19 @@ export async function generateMoodStatusForUser(
   options?: {
     locale?: string | null;
     force?: boolean;
+    /** v1.8.3 — read-only navigation path; see weight-status for the rationale. */
+    readOnly?: boolean;
   },
 ): Promise<{
   hasProvider: boolean;
   text: string | null;
   cached: boolean;
   updatedAt: string | null;
+  preparing?: boolean;
 }> {
   const locale = normalizeLocale(options?.locale);
   const force = options?.force === true;
+  const readOnly = options?.readOnly === true;
   const cacheAction = `insights.mood-status.${locale}`;
   const todayKey = toBerlinDayKey(new Date());
 
@@ -97,6 +104,29 @@ export async function generateMoodStatusForUser(
     };
   }
 
+  if (readOnly) {
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId,
+      metric: "mood",
+      locale,
+    });
+    if (outcome === "no-provider") {
+      return {
+        hasProvider: false,
+        text: getNoKeyMoodStatusText(locale),
+        cached: true,
+        updatedAt: null,
+      };
+    }
+    return {
+      hasProvider: true,
+      text: null,
+      cached: false,
+      updatedAt: null,
+      preparing: true,
+    };
+  }
+
   // v1.4.28 FB-D2 — cap at 90 entries (~3 months) for the mood
   // surface. The downstream `applyPayloadBudget` slims further.
   // Order desc + reverse so the per-row consumers downstream (which
@@ -382,6 +412,8 @@ export async function generateMoodStatusForUser(
     return returnTimeoutFallback({
       cacheAction,
       reason: outcome.kind,
+      userId,
+      todayKey,
       stubText: getNoKeyMoodStatusText(locale),
     });
   }

src/lib/insights/pulse-status.ts

@@ -28,7 +28,10 @@ import {
   summarizeSeries,
 } from "@/lib/insights/status-shared";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
-import { readFreshStatusText } from "@/lib/insights/status-cache";
+import {
+  readFreshStatusText,
+  resolveReadOnlyStatusMiss,
+} from "@/lib/insights/status-cache";
 import { returnTimeoutFallback } from "@/lib/insights/timeout-fallback";
 import { annotate } from "@/lib/logging/context";
 import { toBerlinDayKey } from "@/lib/tz/resolver";
@@ -38,15 +41,19 @@ export async function generatePulseStatusForUser(
   options?: {
     locale?: string | null;
     force?: boolean;
+    /** v1.8.3 — read-only navigation path; see weight-status for the rationale. */
+    readOnly?: boolean;
   },
 ): Promise<{
   hasProvider: boolean;
   text: string | null;
   cached: boolean;
   updatedAt: string | null;
+  preparing?: boolean;
 }> {
   const locale = normalizeLocale(options?.locale);
   const force = options?.force === true;
+  const readOnly = options?.readOnly === true;
   const cacheAction = `insights.pulse-status.${locale}`;
   const todayKey = toBerlinDayKey(new Date());
 
@@ -65,6 +72,29 @@ export async function generatePulseStatusForUser(
     };
   }
 
+  if (readOnly) {
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId,
+      metric: "pulse",
+      locale,
+    });
+    if (outcome === "no-provider") {
+      return {
+        hasProvider: false,
+        text: getNoKeyPulseStatusText(locale),
+        cached: true,
+        updatedAt: null,
+      };
+    }
+    return {
+      hasProvider: true,
+      text: null,
+      cached: false,
+      updatedAt: null,
+      preparing: true,
+    };
+  }
+
   const user = await prisma.user.findUnique({
     where: { id: userId },
     select: {
@@ -270,6 +300,8 @@ export async function generatePulseStatusForUser(
     return returnTimeoutFallback({
       cacheAction,
       reason: outcome.kind,
+      userId,
+      todayKey,
       stubText: getNoKeyPulseStatusText(locale),
     });
   }

src/lib/insights/status-cache.ts

@@ -1,4 +1,10 @@
 import { prisma } from "@/lib/db";
+import { hasUsableStatusProvider } from "@/lib/insights/status-provider";
+import {
+  enqueueStatusGeneration,
+  type InsightStatusMetric,
+} from "@/lib/jobs/insight-status-generate-shared";
+import { annotate } from "@/lib/logging/context";
 
 /**
  * Shared cache-read for the seven `*-status.ts` insight generators.
@@ -31,6 +37,8 @@ interface ParsedStatusCache {
   model?: string;
   tokensUsed?: number | null;
   timeout?: boolean;
+  /** v1.8.3 — ISO timestamp before which a timeout stub suppresses re-enqueue. */
+  retryAt?: string;
 }
 
 /**
@@ -92,3 +100,91 @@ export async function readFreshStatusText(args: {
     return null;
   }
 }
+
+/**
+ * Outcome of the read-only cache-miss resolution. The generators map this
+ * onto their public return shape:
+ *   - `no-provider` → `{ hasProvider: false, text: <no-key fallback> }`
+ *   - `preparing`   → `{ hasProvider: true, text: null, preparing: true }`
+ */
+export type ReadOnlyMissOutcome = "no-provider" | "preparing";
+
+/**
+ * v1.8.3 — resolve what a read-only status generation should return on a
+ * cache miss WITHOUT running the heavy SQL gather or the blocking LLM
+ * round-trip. This is the core of the holistic freeze fix: a navigation
+ * request must never await an uncapped provider call.
+ *
+ * On a miss the route either:
+ *   - finds no usable provider → returns `no-provider` (the card shows the
+ *     no-key fallback; nothing to generate), or
+ *   - finds a provider → fire-and-forget enqueues a generation job and
+ *     returns `preparing` (the card shows a preparing state and the client
+ *     polls until the worker warms the cache).
+ *
+ * The provider probe is a cheap chain-resolve (no completion), so the GET
+ * stays sub-second even on a cold cache.
+ */
+export async function resolveReadOnlyStatusMiss(args: {
+  userId: string;
+  metric: InsightStatusMetric;
+  locale: "de" | "en";
+}): Promise<ReadOnlyMissOutcome> {
+  const hasProvider = await hasUsableStatusProvider(args.userId);
+  if (!hasProvider) return "no-provider";
+
+  const cacheAction = `insights.${args.metric}-status.${args.locale}`;
+
+  // v1.8.3 — honour the short-TTL negative cache. If the worker recently
+  // hit a provider stall it wrote a `retryAt` stub; re-enqueuing on every
+  // navigation while the provider is still degraded would be a storm. While
+  // the stub is fresh, stay in `preparing` without enqueuing; once it goes
+  // stale (or never existed) enqueue a fresh generation.
+  if (await hasFreshTimeoutStub({ userId: args.userId, cacheAction })) {
+    annotate({
+      action: { name: "insights.status.preparing" },
+      meta: { metric: args.metric, suppressed_enqueue: true },
+    });
+    return "preparing";
+  }
+
+  // Enqueue out of band — do NOT await an LLM here. The enqueue itself is
+  // best-effort and de-duped per (user, metric, locale).
+  await enqueueStatusGeneration({
+    userId: args.userId,
+    metric: args.metric,
+    locale: args.locale,
+  });
+  annotate({
+    action: { name: "insights.status.preparing" },
+    meta: { metric: args.metric },
+  });
+  return "preparing";
+}
+
+/**
+ * True when the most recent cache row for `(userId, cacheAction)` is a
+ * timeout stub whose `retryAt` is still in the future. Used by the
+ * read-only resolver to suppress a re-enqueue storm while a provider is
+ * degraded. A stub without `retryAt` (legacy) is treated as stale so the
+ * resolver retries, matching the pre-v1.8.3 "transient miss" behaviour.
+ */
+async function hasFreshTimeoutStub(args: {
+  userId: string;
+  cacheAction: string;
+}): Promise<boolean> {
+  const latest = await prisma.auditLog.findFirst({
+    where: { userId: args.userId, action: args.cacheAction },
+    orderBy: { createdAt: "desc" },
+    select: { details: true },
+  });
+  if (!latest?.details) return false;
+  try {
+    const parsed = JSON.parse(latest.details) as ParsedStatusCache;
+    if (!isTimeoutStub(parsed)) return false;
+    if (typeof parsed.retryAt !== "string") return false;
+    return new Date(parsed.retryAt).getTime() > Date.now();
+  } catch {
+    return false;
+  }
+}

src/lib/insights/status-provider.ts

@@ -64,6 +64,22 @@ async function resolveStatusChain(
   return [{ providerType: "admin-openai", instance: legacy }];
 }
 
+/**
+ * Cheap provider-availability probe for the read-only status path.
+ *
+ * v1.8.3 — when a status route runs in read-only mode (serve cache, never
+ * block on the LLM) it still has to tell the difference between "no
+ * provider — show the no-key fallback" and "provider configured but the
+ * assessment isn't warm yet — show preparing + enqueue a generation". This
+ * resolves the same chain `runStatusCompletion` would, but does NOT run a
+ * completion, so the navigation request never awaits an LLM round-trip.
+ */
+export async function hasUsableStatusProvider(
+  userId: string,
+): Promise<boolean> {
+  return (await resolveStatusChain(userId)) !== null;
+}
+
 /**
  * Run a status generation across the user's provider chain, bounded by
  * `STATUS_PROVIDER_TIMEOUT_MS`. The result discriminates between

src/lib/insights/timeout-fallback.ts

@@ -1,32 +1,79 @@
+import { prisma } from "@/lib/db";
 import { annotate } from "@/lib/logging/context";
 
 /**
  * Shared timeout / provider-error fallback for the InsightStatus routes.
  *
  * When the upstream provider call exceeds `STATUS_PROVIDER_TIMEOUT_MS`
- * (or rejects outright) the route still has to return a deterministic
- * envelope so the InsightStatusCard renders the no-key fallback text
- * instead of spinning. It used to ALSO persist a sentinel `auditLog`
- * row keyed to today's Berlin day. That stub carried the generic
- * fallback under the same `text` field a real assessment uses, the
- * cache-read served it as valid, and it stuck until midnight — hiding
- * the real, data-driven assessment for the rest of the day on any
- * account that hit a single transient stall.
+ * (or rejects outright) the generator still has to return a deterministic
+ * envelope. It used to persist a sentinel row carrying the generic
+ * fallback under the same `text` field a real assessment uses; the
+ * cache-read served it as valid and it stuck until midnight — hiding the
+ * real, data-driven assessment for the rest of the day on any account that
+ * hit a single transient stall. v1.4.28 removed that persist entirely.
  *
- * The fix removes the persist entirely: a timeout / error is treated as
- * a transient miss. The current render still gets the fallback text, but
- * nothing is written, so the next mount re-attempts a real generation.
- * The shared cache-read (`status-cache.ts`) additionally rejects any
- * legacy stub rows still in the table by their `model:"timeout-stub"` /
- * `timeout:true` markers.
- *
- * `cached:true` is kept so the UI does not mislabel the fallback as a
- * fresh assessment, and `updatedAt:null` signals "no persisted row".
+ * v1.8.3 — re-introduce a *negative* cache, but a strictly bounded one. The
+ * status route is now read-only: a cache miss enqueues an out-of-band
+ * generation. If the provider stalls inside the worker, nothing stops the
+ * next navigation from enqueuing again, and again — a stalled provider
+ * turns into a re-enqueue storm. So the timeout path writes a short-TTL
+ * negative stub (`{ timeout:true, model:"timeout-stub", retryAt }`). The
+ * read-only resolver honours it: while the stub is fresh it returns
+ * `preparing` WITHOUT re-enqueuing; once `retryAt` passes the stub is stale
+ * and the next visit re-attempts. The stub is NEVER served as assessment
+ * text — `readFreshStatusText` still rejects every timeout stub by marker,
+ * so a transient stall can't hide the real assessment for the day.
+ */
+
+/** Short retry window after a provider stall before a re-enqueue is allowed. */
+export const TIMEOUT_NEGATIVE_CACHE_MS = 5 * 60 * 1000;
+
+/**
+ * Persist a short-TTL negative-cache stub for `(userId, cacheAction)`.
+ * Best-effort: a failure here just means the read-only resolver may
+ * re-enqueue sooner than the window intended.
+ */
+export async function persistTimeoutNegativeStub(args: {
+  userId: string;
+  cacheAction: string;
+  todayKey: string;
+  reason: "timeout" | "error";
+}): Promise<void> {
+  try {
+    await prisma.auditLog.create({
+      data: {
+        userId: args.userId,
+        action: args.cacheAction,
+        details: JSON.stringify({
+          dateKey: args.todayKey,
+          timeout: true,
+          model: "timeout-stub",
+          reason: args.reason,
+          retryAt: new Date(
+            Date.now() + TIMEOUT_NEGATIVE_CACHE_MS,
+          ).toISOString(),
+        }),
+      },
+    });
+  } catch {
+    // Negative cache is best-effort — never throw out of the timeout path.
+  }
+}
+
+/**
+ * Return the deterministic fallback envelope for a timeout / error. The
+ * current render still gets the fallback text, marked `cached:true` so the
+ * UI does not mislabel it as a fresh assessment and `updatedAt:null` to
+ * signal "no persisted assessment". When `userId` + `todayKey` are present
+ * (the worker-side generation path) a short-TTL negative stub is persisted
+ * so a stalled provider doesn't trigger a re-enqueue storm.
  */
 export function returnTimeoutFallback(input: {
   cacheAction: string;
   reason: "timeout" | "error";
   stubText: string;
+  userId?: string;
+  todayKey?: string;
 }): {
   hasProvider: true;
   text: string;
@@ -37,6 +84,15 @@ export function returnTimeoutFallback(input: {
     action: { name: "insights.status.fallback_served" },
     meta: { cacheAction: input.cacheAction, reason: input.reason },
   });
+  if (input.userId && input.todayKey) {
+    // Fire-and-forget — the caller's render must not wait on the write.
+    void persistTimeoutNegativeStub({
+      userId: input.userId,
+      cacheAction: input.cacheAction,
+      todayKey: input.todayKey,
+      reason: input.reason,
+    });
+  }
   return {
     hasProvider: true,
     text: input.stubText,

src/lib/insights/weight-status.ts

@@ -32,7 +32,10 @@ import {
   summarizeSeries,
 } from "@/lib/insights/status-shared";
 import { runStatusCompletion } from "@/lib/insights/status-provider";
-import { readFreshStatusText } from "@/lib/insights/status-cache";
+import {
+  readFreshStatusText,
+  resolveReadOnlyStatusMiss,
+} from "@/lib/insights/status-cache";
 import { returnTimeoutFallback } from "@/lib/insights/timeout-fallback";
 import { annotate } from "@/lib/logging/context";
 import { toBerlinDayKey } from "@/lib/tz/resolver";
@@ -81,15 +84,25 @@ export async function generateWeightStatusForUser(
   options?: {
     locale?: string | null;
     force?: boolean;
+    /**
+     * v1.8.3 — read-only mode for the navigation path. On a cache miss the
+     * generator does NOT run the SQL gather + blocking LLM inline; it
+     * enqueues an out-of-band generation and returns a `preparing` shape so
+     * the request never awaits the provider. The worker (and the nightly
+     * pre-generate cron) call with `readOnly: false` to actually generate.
+     */
+    readOnly?: boolean;
   },
 ): Promise<{
   hasProvider: boolean;
   text: string | null;
   cached: boolean;
   updatedAt: string | null;
+  preparing?: boolean;
 }> {
   const locale = normalizeLocale(options?.locale);
   const force = options?.force === true;
+  const readOnly = options?.readOnly === true;
   const cacheAction = `insights.weight-status.${locale}`;
   const todayKey = toBerlinDayKey(new Date());
 
@@ -111,6 +124,31 @@ export async function generateWeightStatusForUser(
     };
   }
 
+  // v1.8.3 — read-only navigation path: never block on the provider.
+  // Enqueue generation out of band and return preparing / no-provider.
+  if (readOnly) {
+    const outcome = await resolveReadOnlyStatusMiss({
+      userId,
+      metric: "weight",
+      locale,
+    });
+    if (outcome === "no-provider") {
+      return {
+        hasProvider: false,
+        text: getNoKeyWeightStatusText(locale),
+        cached: true,
+        updatedAt: null,
+      };
+    }
+    return {
+      hasProvider: true,
+      text: null,
+      cached: false,
+      updatedAt: null,
+      preparing: true,
+    };
+  }
+
   // v1.4.28 FB-D2 — cap the snapshot input. BP captures three types
   // per day so 1095 = 365 d × 3 channels. Order desc + reverse so we
   // keep the most recent year of rows.
@@ -414,6 +452,8 @@ export async function generateWeightStatusForUser(
     return returnTimeoutFallback({
       cacheAction,
       reason: outcome.kind,
+      userId,
+      todayKey,
       stubText: getNoKeyWeightStatusText(locale),
     });
   }

src/lib/jobs/__tests__/insight-pregenerate.test.ts

@@ -209,7 +209,7 @@ describe("runInsightPregenerate — per-metric warm pass", () => {
     return vi.fn().mockResolvedValue(result);
   }
 
-  it("forces all seven status generators after a successful comprehensive generation", async () => {
+  it("forces all seven status generators in BOTH locales after a successful comprehensive generation", async () => {
     const { prisma } = makePrisma([{ id: "u1", locale: "de" }]);
     const generate = vi
       .fn()
@@ -223,13 +223,34 @@ describe("runInsightPregenerate — per-metric warm pass", () => {
       statusGenerators,
     });
 
-    // Every status generator forced with the resolved locale.
+    // v1.8.3 — each generator is forced once per supported locale (de + en)
+    // so the cache the client reads against (active UI locale, not the
+    // persisted User.locale) is always warm.
     for (const g of statusGenerators) {
-      expect(g).toHaveBeenCalledTimes(1);
+      expect(g).toHaveBeenCalledTimes(2);
       expect(g).toHaveBeenCalledWith("u1", { locale: "de", force: true });
+      expect(g).toHaveBeenCalledWith("u1", { locale: "en", force: true });
     }
-    // Seven fresh, provider-backed assessments warmed.
-    expect(result.assessmentsWarmed).toBe(7);
+    // Seven generators × two locales = 14 fresh, provider-backed assessments.
+    expect(result.assessmentsWarmed).toBe(14);
+  });
+
+  it("warms when the comprehensive pass returned `cached` (its write still evicted the per-status caches)", async () => {
+    const { prisma } = makePrisma([{ id: "u1", locale: "de" }]);
+    const generate = vi.fn().mockResolvedValue({ status: "cached" });
+    const statusGenerators = Array.from({ length: 7 }, () =>
+      warmGen({ hasProvider: true, cached: false }),
+    );
+
+    const result = await runInsightPregenerate(prisma as never, {
+      generate,
+      statusGenerators,
+    });
+
+    for (const g of statusGenerators) {
+      expect(g).toHaveBeenCalledTimes(2);
+    }
+    expect(result.assessmentsWarmed).toBe(14);
   });
 
   it("does NOT warm when the comprehensive pass skipped (no provider)", async () => {
@@ -290,11 +311,13 @@ describe("runInsightPregenerate — per-metric warm pass", () => {
       statusGenerators,
     });
 
-    // The throw must not abort the loop — every generator was attempted.
+    // The throw must not abort the loop — every generator was attempted
+    // once per supported locale (de + en).
     for (const g of statusGenerators) {
-      expect(g).toHaveBeenCalledTimes(1);
+      expect(g).toHaveBeenCalledTimes(2);
     }
-    expect(result.assessmentsWarmed).toBe(2);
+    // Two counting generators × two locales = 4 fresh assessments.
+    expect(result.assessmentsWarmed).toBe(4);
   });
 });
 

src/lib/jobs/__tests__/insight-status-generate.test.ts

@@ -0,0 +1,156 @@
+/**
+ * v1.8.3 — unit tests for the on-demand per-metric status generation queue.
+ *
+ * Covers:
+ *   - the dispatch runs the matching generator with force:true;
+ *   - an unknown metric is annotated and skipped, not thrown;
+ *   - the enqueue helper no-ops without a global boss instance and de-dupes
+ *     with a singletonKey when one is present;
+ *   - the queue is registered in `allQueues` AND wired to a `boss.work`
+ *     handler in reminder-worker.ts (the v1.4.37 W10 unregistered-queue
+ *     catch — a queue built but never registered silently never drains).
+ */
+import { describe, it, expect, vi, beforeEach } from "vitest";
+import fs from "node:fs";
+import path from "node:path";
+
+const getGlobalBoss = vi.fn();
+vi.mock("@/lib/jobs/boss-instance", () => ({
+  getGlobalBoss: (...a: unknown[]) => getGlobalBoss(...a),
+}));
+// The dispatch module imports the seven generators transitively; stub them
+// so the test never reaches a live provider.
+vi.mock("@/lib/insights/general-status", () => ({
+  generateGeneralStatusForUser: vi.fn(),
+}));
+vi.mock("@/lib/insights/blood-pressure-status", () => ({
+  generateBloodPressureStatusForUser: vi.fn(),
+}));
+vi.mock("@/lib/insights/weight-status", () => ({
+  generateWeightStatusForUser: vi.fn(),
+}));
+vi.mock("@/lib/insights/pulse-status", () => ({
+  generatePulseStatusForUser: vi.fn(),
+}));
+vi.mock("@/lib/insights/bmi-status", () => ({
+  generateBmiStatusForUser: vi.fn(),
+}));
+vi.mock("@/lib/insights/mood-status", () => ({
+  generateMoodStatusForUser: vi.fn(),
+}));
+vi.mock("@/lib/insights/medication-compliance-status", () => ({
+  generateMedicationComplianceStatusForUser: vi.fn(),
+}));
+
+import {
+  runInsightStatusGenerate,
+  enqueueStatusGeneration,
+  INSIGHT_STATUS_GENERATE_QUEUE,
+  INSIGHT_STATUS_METRICS,
+} from "../insight-status-generate";
+
+beforeEach(() => {
+  vi.clearAllMocks();
+});
+
+describe("runInsightStatusGenerate", () => {
+  it("forces the matching generator with the payload locale", async () => {
+    const weight = vi.fn().mockResolvedValue(undefined);
+    const generators = {
+      general: vi.fn(),
+      "blood-pressure": vi.fn(),
+      weight,
+      pulse: vi.fn(),
+      bmi: vi.fn(),
+      mood: vi.fn(),
+      "medication-compliance": vi.fn(),
+    };
+    await runInsightStatusGenerate(
+      { userId: "u1", metric: "weight", locale: "en" },
+      generators,
+    );
+    expect(weight).toHaveBeenCalledWith("u1", { locale: "en", force: true });
+    // No other generator ran.
+    expect(generators.pulse).not.toHaveBeenCalled();
+  });
+
+  it("skips (does not throw) an unknown metric", async () => {
+    await expect(
+      runInsightStatusGenerate(
+        // eslint-disable-next-line @typescript-eslint/no-explicit-any
+        { userId: "u1", metric: "nope" as any, locale: "de" },
+        // eslint-disable-next-line @typescript-eslint/no-explicit-any
+        {} as any,
+      ),
+    ).resolves.toBeUndefined();
+  });
+
+  it("exposes the seven canonical metrics", () => {
+    expect([...INSIGHT_STATUS_METRICS].sort()).toEqual(
+      [
+        "general",
+        "blood-pressure",
+        "weight",
+        "pulse",
+        "bmi",
+        "mood",
+        "medication-compliance",
+      ].sort(),
+    );
+  });
+});
+
+describe("enqueueStatusGeneration", () => {
+  it("no-ops when no global boss is available", async () => {
+    getGlobalBoss.mockReturnValue(null);
+    await expect(
+      enqueueStatusGeneration({ userId: "u1", metric: "weight", locale: "de" }),
+    ).resolves.toBeUndefined();
+  });
+
+  it("sends with a per-(user,metric,locale) singletonKey when boss exists", async () => {
+    const send = vi.fn().mockResolvedValue("job-id");
+    getGlobalBoss.mockReturnValue({ send });
+    await enqueueStatusGeneration({
+      userId: "u1",
+      metric: "mood",
+      locale: "en",
+    });
+    expect(send).toHaveBeenCalledWith(
+      INSIGHT_STATUS_GENERATE_QUEUE,
+      { userId: "u1", metric: "mood", locale: "en" },
+      expect.objectContaining({ singletonKey: "u1:mood:en" }),
+    );
+  });
+
+  it("swallows a boss.send rejection (best-effort enqueue)", async () => {
+    const send = vi.fn().mockRejectedValue(new Error("boom"));
+    getGlobalBoss.mockReturnValue({ send });
+    await expect(
+      enqueueStatusGeneration({ userId: "u1", metric: "bmi", locale: "de" }),
+    ).resolves.toBeUndefined();
+  });
+});
+
+describe("queue registration", () => {
+  const workerSrc = fs.readFileSync(
+    path.resolve(__dirname, "../reminder-worker.ts"),
+    "utf8",
+  );
+
+  it("registers the queue name in the allQueues createQueue loop", () => {
+    const match = workerSrc.match(/const allQueues\s*=\s*\[([\s\S]*?)\];/);
+    expect(match).not.toBeNull();
+    expect(match![1]).toMatch(/\bINSIGHT_STATUS_GENERATE_QUEUE\b/);
+  });
+
+  it("registers a boss.work handler for the queue", () => {
+    expect(workerSrc).toMatch(
+      /boss\.work[\s\S]{0,120}INSIGHT_STATUS_GENERATE_QUEUE/,
+    );
+  });
+
+  it("exposes a stable queue name", () => {
+    expect(INSIGHT_STATUS_GENERATE_QUEUE).toBe("insight-status-generate");
+  });
+});

src/lib/jobs/insight-pregenerate.ts

@@ -156,12 +156,20 @@ const DEFAULT_STATUS_GENERATORS: ReadonlyArray<StatusGenerator> = [
  */
 async function warmPerStatusCaches(
   userId: string,
-  locale: "de" | "en",
+  locales: ReadonlyArray<"de" | "en">,
   generators: ReadonlyArray<StatusGenerator>,
 ): Promise<number> {
   const limit = pLimit(WARM_PASS_CONCURRENCY);
-  const results = await Promise.all(
-    generators.map((generate) =>
+  // v1.8.3 — warm every locale the user might request, not just the
+  // User.locale. The per-status cache key is
+  // `insights.<metric>-status.<locale>`, but the client sends the active
+  // UI-cookie / Accept-Language locale (narrowed to de|en), which can
+  // differ from the persisted User.locale (e.g. null → "de" in the cron
+  // while the browser cookie is "en"). Warming only the User.locale left
+  // the locale the client actually reads cold, so the first visit still
+  // ran a live generation per card — the exact freeze this release closes.
+  const tasks = generators.flatMap((generate) =>
+    locales.map((locale) =>
       limit(async () => {
         try {
           const outcome = await generate(userId, { locale, force: true });
@@ -174,6 +182,7 @@ async function warmPerStatusCaches(
       }),
     ),
   );
+  const results = await Promise.all(tasks);
   return results.reduce((sum: number, n) => sum + n, 0);
 }
 
@@ -301,18 +310,23 @@ export async function runInsightPregenerate(
     }
 
     // Warm the per-metric assessment caches the comprehensive generator
-    // evicts (`evictPerStatusInsightCache`) but does not re-fill. Only
-    // do this when the comprehensive pass actually produced an insight:
-    // a `skipped` (no-provider) user has nothing to warm, and a `failed`
-    // pass means the provider chain is unhealthy this run — re-running
-    // it seven more times would only multiply the failure. A `cached`
-    // outcome can't happen here (we always force), but if a future
-    // change reintroduces it the per-metric caches were not evicted, so
-    // there is nothing to warm.
-    if (outcome.status === "generated") {
+    // evicts (`evictPerStatusInsightCache`) but does not re-fill.
+    //
+    // v1.8.3 — warm on `generated` OR `cached`. A `cached` outcome means
+    // the comprehensive insight is fresh-enough that its write already
+    // evicted the per-status caches, so they are cold and need re-filling
+    // exactly as on a fresh `generated`. The pre-v1.8.3 `=== "generated"`
+    // gate left those caches cold whenever the comprehensive pass short-
+    // circuited to `cached`, so the morning visit still ran a live
+    // generation per card. `skipped` (no provider) has nothing to warm and
+    // `failed` means the chain is unhealthy this run — re-running it seven
+    // more times would only multiply the failure.
+    if (outcome.status === "generated" || outcome.status === "cached") {
       result.assessmentsWarmed += await warmPerStatusCaches(
         candidate.id,
-        locale,
+        // Warm both supported locales — the client reads against the active
+        // UI locale, which can differ from the persisted User.locale.
+        ["de", "en"],
         statusGenerators,
       );
     }

src/lib/jobs/insight-status-generate-shared.ts

@@ -0,0 +1,91 @@
+/**
+ * v1.8.3 — shared contract for the on-demand per-metric status generation
+ * queue, free of any generator import.
+ *
+ * The status generators import `enqueueStatusGeneration` from here (via
+ * `status-cache.ts`). Keeping the enqueue + the queue name + the metric
+ * vocabulary in a generator-free module breaks the import cycle that would
+ * otherwise form: generator → status-cache → (enqueue) → generator. The
+ * worker-only dispatch (`runInsightStatusGenerate`, which references the
+ * concrete generators) lives in `insight-status-generate.ts` and is
+ * imported solely by the worker boot file.
+ */
+import { getGlobalBoss } from "@/lib/jobs/boss-instance";
+import { annotate } from "@/lib/logging/context";
+
+export const INSIGHT_STATUS_GENERATE_QUEUE = "insight-status-generate";
+
+/**
+ * Low concurrency keeps the on-demand warm pass from saturating the
+ * Prisma pool or fanning out an unbounded number of concurrent provider
+ * calls when several cards request generation on the same first visit.
+ */
+export const INSIGHT_STATUS_GENERATE_CONCURRENCY = 2;
+
+/**
+ * The seven per-metric status scopes the queue can generate. Mirrors the
+ * `cacheAction` slug each generator uses (`insights.<metric>-status.<locale>`)
+ * so the queue payload, the route, and the client all speak one vocabulary.
+ */
+export type InsightStatusMetric =
+  | "general"
+  | "blood-pressure"
+  | "weight"
+  | "pulse"
+  | "bmi"
+  | "mood"
+  | "medication-compliance";
+
+export const INSIGHT_STATUS_METRICS: ReadonlyArray<InsightStatusMetric> = [
+  "general",
+  "blood-pressure",
+  "weight",
+  "pulse",
+  "bmi",
+  "mood",
+  "medication-compliance",
+];
+
+export interface InsightStatusGeneratePayload {
+  userId: string;
+  metric: InsightStatusMetric;
+  /** Resolved client locale — the cache key the route reads against. */
+  locale: "de" | "en";
+}
+
+/**
+ * Fire-and-forget enqueue used by the read-only status route on a cache
+ * miss. A `singletonKey` per `(user, metric, locale)` collapses repeated
+ * polls into one queued job, so a client polling every few seconds while
+ * the provider works can't pile up duplicate generations. No-ops cleanly
+ * when the global boss instance is not available (e.g. the web process
+ * runs without an embedded worker) — the nightly pre-generate cron and
+ * the daily status crons remain the catch-net.
+ */
+export async function enqueueStatusGeneration(
+  payload: InsightStatusGeneratePayload,
+): Promise<void> {
+  const boss = getGlobalBoss();
+  if (!boss) {
+    annotate({
+      action: { name: "insights.status.generate.no_boss" },
+      meta: { metric: payload.metric },
+    });
+    return;
+  }
+  try {
+    await boss.send(INSIGHT_STATUS_GENERATE_QUEUE, payload, {
+      singletonKey: `${payload.userId}:${payload.metric}:${payload.locale}`,
+      // De-dupe within a short window so a polling client doesn't enqueue
+      // a fresh job on every tick while a generation is already running.
+      singletonSeconds: 120,
+    });
+    annotate({
+      action: { name: "insights.status.generate.enqueued" },
+      meta: { metric: payload.metric },
+    });
+  } catch {
+    // Enqueue is best-effort; a failure here just means the card stays
+    // in `preparing` until the next poll / nightly cron warms the cache.
+  }
+}

src/lib/jobs/insight-status-generate.ts

@@ -0,0 +1,92 @@
+/**
+ * v1.8.3 — on-demand per-metric status generation queue (worker dispatch).
+ *
+ * Before this, every `/insights/<metric>` sub-page fetched its status
+ * card from `GET /api/insights/<metric>-status`, and that GET ran the
+ * full SQL gather + a blocking LLM round-trip inline. The provider call
+ * was capped server-side at 60 s, so a cold cache on a slow provider
+ * pinned the navigation request — and, because the timeout fallback is
+ * deliberately not persisted, every subsequent visit re-blocked.
+ *
+ * This queue moves the heavy gather+LLM off the request path. The route
+ * now reads the cache only; on a miss it fire-and-forget enqueues a job
+ * (via `enqueueStatusGeneration` in the generator-free shared module) and
+ * returns a `preparing` envelope. The worker runs the matching generator
+ * with `force: true`, which persists the assessment cache row; the client
+ * polls the read-only GET until it lands.
+ *
+ * Recurring/event pg-boss task — never runs inside an HTTP request and
+ * never shells out to `tsx` (CLAUDE.md DO-NOTs). Registered in the
+ * reminder-worker `allQueues` loop + a `boss.work` handler; a queue built
+ * but never registered silently never drains (the v1.4.37 W10 incident),
+ * so a guard test pins the registration.
+ *
+ * The queue name, payload type, metric vocabulary, and the enqueue helper
+ * live in `insight-status-generate-shared.ts` so the status generators can
+ * import the enqueue without pulling the concrete generators back in (an
+ * import cycle). This module re-exports them for the worker's convenience.
+ */
+import { annotate } from "@/lib/logging/context";
+import { generateGeneralStatusForUser } from "@/lib/insights/general-status";
+import { generateBloodPressureStatusForUser } from "@/lib/insights/blood-pressure-status";
+import { generateWeightStatusForUser } from "@/lib/insights/weight-status";
+import { generatePulseStatusForUser } from "@/lib/insights/pulse-status";
+import { generateBmiStatusForUser } from "@/lib/insights/bmi-status";
+import { generateMoodStatusForUser } from "@/lib/insights/mood-status";
+import { generateMedicationComplianceStatusForUser } from "@/lib/insights/medication-compliance-status";
+import {
+  INSIGHT_STATUS_GENERATE_QUEUE,
+  INSIGHT_STATUS_GENERATE_CONCURRENCY,
+  INSIGHT_STATUS_METRICS,
+  enqueueStatusGeneration,
+  type InsightStatusMetric,
+  type InsightStatusGeneratePayload,
+} from "@/lib/jobs/insight-status-generate-shared";
+
+export {
+  INSIGHT_STATUS_GENERATE_QUEUE,
+  INSIGHT_STATUS_GENERATE_CONCURRENCY,
+  INSIGHT_STATUS_METRICS,
+  enqueueStatusGeneration,
+};
+export type { InsightStatusMetric, InsightStatusGeneratePayload };
+
+/**
+ * Narrow generator shape: the queue only forces a fresh generation and
+ * discards the payload (the generator persists the cache row itself).
+ */
+type StatusGenerator = (
+  userId: string,
+  options: { locale: "de" | "en"; force: boolean },
+) => Promise<unknown>;
+
+const GENERATORS: Record<InsightStatusMetric, StatusGenerator> = {
+  general: generateGeneralStatusForUser,
+  "blood-pressure": generateBloodPressureStatusForUser,
+  weight: generateWeightStatusForUser,
+  pulse: generatePulseStatusForUser,
+  bmi: generateBmiStatusForUser,
+  mood: generateMoodStatusForUser,
+  "medication-compliance": generateMedicationComplianceStatusForUser,
+};
+
+/**
+ * Run one on-demand status generation. Pure of pg-boss so the unit test
+ * can drive it directly. Forces a fresh generation so the cache row the
+ * route's read-only path missed gets written; an unknown metric is
+ * annotated and skipped rather than thrown.
+ */
+export async function runInsightStatusGenerate(
+  payload: InsightStatusGeneratePayload,
+  generators: Record<InsightStatusMetric, StatusGenerator> = GENERATORS,
+): Promise<void> {
+  const generate = generators[payload.metric];
+  if (!generate) {
+    annotate({
+      action: { name: "insights.status.generate.unknown_metric" },
+      meta: { metric: payload.metric },
+    });
+    return;
+  }
+  await generate(payload.userId, { locale: payload.locale, force: true });
+}

src/lib/jobs/reminder-worker.ts

@@ -67,6 +67,12 @@ import {
   runInsightPregenerate,
   type InsightPregeneratePayload,
 } from "@/lib/jobs/insight-pregenerate";
+import {
+  INSIGHT_STATUS_GENERATE_QUEUE,
+  INSIGHT_STATUS_GENERATE_CONCURRENCY,
+  runInsightStatusGenerate,
+  type InsightStatusGeneratePayload,
+} from "@/lib/jobs/insight-status-generate";
 import {
   INTAKE_AUTO_SKIP_QUEUE,
   INTAKE_AUTO_SKIP_CRON,
@@ -1700,6 +1706,32 @@ async function handleInsightPregenerateJob(
   });
 }
 
+/**
+ * v1.8.3 — on-demand per-metric status generation. The read-only status
+ * route enqueues one job per cold card; this handler runs the matching
+ * generator with `force: true` so the assessment cache row lands and the
+ * polling client picks it up. Each job carries `{ userId, metric, locale }`.
+ */
+async function handleInsightStatusGenerate(
+  jobs: Job<InsightStatusGeneratePayload>[],
+) {
+  await withBackgroundEvent("job.insight_status_generate", async (evt) => {
+    for (const job of jobs) {
+      if (!job.data?.userId || !job.data?.metric) continue;
+      try {
+        await runInsightStatusGenerate(job.data);
+        evt.addMeta("status_generated", `${job.data.metric}:${job.data.locale}`);
+      } catch (err) {
+        evt.addWarning(
+          `insight-status-generate failed for ${job.data.metric}: ${
+            err instanceof Error ? err.message : String(err)
+          }`,
+        );
+      }
+    }
+  });
+}
+
 /**
  * v1.4.46 — hourly auto-skip for stale unmarked medication intakes.
  *
@@ -2055,6 +2087,12 @@ export async function startReminderWorker() {
     // this entry the 04:30 schedule silently no-ops and every briefing
     // falls back to the lazy on-demand generation it was meant to retire.
     INSIGHT_PREGENERATE_QUEUE,
+    // v1.8.3 — on-demand per-metric status generation. The read-only
+    // status route enqueues here on a cold card; without this entry
+    // pg-boss never provisions the queue and the enqueue silently drops,
+    // leaving every status card stuck on "preparing". No cron schedule —
+    // it is a send-only queue driven by navigation.
+    INSIGHT_STATUS_GENERATE_QUEUE,
     // v1.7.0 — soft-deleted measurement tombstone prune. Same createQueue
     // contract as the other cleanup jobs; without this entry the daily
     // schedule silently no-ops and pruned-past-retention tombstones pile
@@ -2306,6 +2344,15 @@ export async function startReminderWorker() {
     { localConcurrency: 1 },
     handleInsightPregenerateJob,
   );
+  // v1.8.3 — on-demand per-metric status generation enqueued by the
+  // read-only status route on a cold card. Low concurrency so a first
+  // visit that cold-misses several cards can't saturate the Prisma pool
+  // or fan out an unbounded number of concurrent provider calls.
+  await boss.work<InsightStatusGeneratePayload>(
+    INSIGHT_STATUS_GENERATE_QUEUE,
+    { localConcurrency: INSIGHT_STATUS_GENERATE_CONCURRENCY },
+    handleInsightStatusGenerate,
+  );
   // v1.4.46 — hourly auto-skip for stale unmarked intakes.
   // Single-flight: two ticks racing against the same row pile is wasted
   // work, and the underlying `updateMany` is the canonical idempotent