test(webapp): assert the card and the text report stay one report

Landmarks derived from the layout spec must appear in the same order in the server-rendered card
and in the markdown, over healthy, degraded, untrustworthy and empty reports.

Author: kathiekiwi

apps/webapp/test/__snapshots__/reportRenderParity.test.ts.snap

@@ -0,0 +1,161 @@
+// Vitest Snapshot v1, https://vitest.dev/guide/snapshot.html
+
+exports[`ANSI surface > paints the same layout (degraded) 1`] = `
+"/report health        <grey>prod · last 1h · vs your 7d normal</>
+
+<red>✕</> Flow stalled — at your env concurrency limit <amber>for the last 38 min</>
+
+  concurrency     50/50                <green>▁</><green>▂</><green>▃</><green>▄</><green>▅</><amber>▆</><amber>▇</><amber>█</>   38 min at limit
+
+  pending         4,812       <amber>↑ 120×</>   <green>▁</><green>▂</><green>▃</><green>▄</><green>▅</><amber>▆</><amber>▇</><amber>█</>   <grey>(normal ~40)</>
+
+  start latency   p95 43s     <amber>↑ 6×</>     <green>▁</><green>▂</><green>▃</><green>▄</><green>▅</><amber>▆</><amber>▇</><amber>█</>   <grey>(normal ~7s)</>
+
+  throughput      −180/min
+    done          820/min
+    triggered     1,000/min
+
+<grey>  why:  71% of pending is demo-email-sends</>
+<grey>        not your code — failures and durations normal</>
+<grey>        runs are finishing at ~820/min</>
+
+<green>✓</> EXECUTION   the runs that DO start are fine
+
+<green>✓</> LIVENESS    fresh — telemetry current, updated 18s ago
+
+<grey>  read: limit saturated → incoming work exceeds capacity → backlog grows</>
+<grey>        NOT a code problem</>
+
+→ Contact us to raise the limit
+  Read concurrency docs
+  or do nothing — backlog drains in ~26.7 min once triggers ease"
+`;
+
+exports[`ANSI surface > paints the same layout (empty) 1`] = `
+"/report health        <grey>prod · last 1h</>
+
+<green>✓</> health ok
+
+<green>✓</> data fresh
+
+→ nothing to do"
+`;
+
+exports[`ANSI surface > paints the same layout (healthy) 1`] = `
+"/report health        <grey>prod · last 1h · vs your 7d normal</>
+
+<green>✓</> Flow healthy — starting normally
+
+  start latency   p95 6.8s   <grey>→ flat</>   <amber>▇</><amber>▆</><green>▁</><green>▅</><amber>█</><green>▂</><green>▃</><amber>█</>   <grey>(normal ~7s)</>
+
+  pending         34         <grey>→ flat</>   <amber>▇</><amber>▆</><green>▁</><green>▅</><amber>█</><green>▂</><green>▃</><amber>█</>   <grey>(normal ~40)</>
+
+  throughput      +12/min
+    done          842/min
+    triggered     830/min
+
+<green>✓</> EXECUTION   the runs that DO start are fine
+
+<green>✓</> LIVENESS    fresh — telemetry current, updated 21s ago
+
+<grey>  read: runs are starting on time</>
+<grey>        runs are completing normally</>
+
+→ nothing to do"
+`;
+
+exports[`ANSI surface > paints the same layout (untrustworthy) 1`] = `
+"/report health  <amber>⚑</> stale data  <grey>prod · last 1h · vs your 7d normal</>
+
+<red>✕</> Flow unknown — data stale
+
+<amber>⚑</> The telemetry behind this report is stale, so the numbers below are informational only.
+
+<red>✕</> EXECUTION   execution can't be assessed — the telemetry is stale
+
+<red>✕</> LIVENESS    stale — no telemetry in 21m
+
+    liveness   21m
+
+→ Check control plane"
+`;
+
+exports[`markdown surface > renders (degraded) 1`] = `
+"/report health        prod · last 1h · vs your 7d normal
+
+🔴 Flow stalled — at your env concurrency limit for the last 38 min
+
+  concurrency     50/50                ▁▂▃▄▅▆▇█   38 min at limit
+
+  pending         4,812       ↑ 120×   ▁▂▃▄▅▆▇█   (normal ~40)
+
+  start latency   p95 43s     ↑ 6×     ▁▂▃▄▅▆▇█   (normal ~7s)
+
+  throughput      −180/min
+    done          820/min
+    triggered     1,000/min
+
+  why:  71% of pending is demo-email-sends
+        not your code — failures and durations normal
+        runs are finishing at ~820/min
+
+🟢 EXECUTION   the runs that DO start are fine
+
+🟢 LIVENESS    fresh — telemetry current, updated 18s ago
+
+  read: limit saturated → incoming work exceeds capacity → backlog grows
+        NOT a code problem
+
+→ Contact us to raise the limit
+  Read concurrency docs
+  or do nothing — backlog drains in ~26.7 min once triggers ease"
+`;
+
+exports[`markdown surface > renders (empty) 1`] = `
+"/report health        prod · last 1h
+
+🟢 health ok
+
+🟢 data fresh
+
+→ nothing to do"
+`;
+
+exports[`markdown surface > renders (healthy) 1`] = `
+"/report health        prod · last 1h · vs your 7d normal
+
+🟢 Flow healthy — starting normally
+
+  start latency   p95 6.8s   → flat   ▇▆▁▅█▂▃█   (normal ~7s)
+
+  pending         34         → flat   ▇▆▁▅█▂▃█   (normal ~40)
+
+  throughput      +12/min
+    done          842/min
+    triggered     830/min
+
+🟢 EXECUTION   the runs that DO start are fine
+
+🟢 LIVENESS    fresh — telemetry current, updated 21s ago
+
+  read: runs are starting on time
+        runs are completing normally
+
+→ nothing to do"
+`;
+
+exports[`markdown surface > renders (untrustworthy) 1`] = `
+"/report health  🚩 stale data  prod · last 1h · vs your 7d normal
+
+🔴 Flow unknown — data stale
+
+🚩 The telemetry behind this report is stale, so the numbers below are informational only.
+
+🔴 EXECUTION   execution can't be assessed — the telemetry is stale
+
+🔴 LIVENESS    stale — no telemetry in 21m
+
+    liveness   21m
+
+→ Check control plane"
+`;

apps/webapp/test/reportHealth.test.ts

@@ -105,28 +105,24 @@ describe("health cause tree (Golden A — env limit saturation)", () => {
     expect(renderReportMarkdown(vm)).toMatchInlineSnapshot(`
       "/report health        prod · last 1h · vs your 7d normal
 
-      🟡 Flow slowing  ·  🟢 Execution healthy  ·  🟢 data fresh
-
-      FLOW        🟡 at your env concurrency limit (last 40 min)
+      🟡 Flow slowing — at your env concurrency limit for the last 40 min
 
         concurrency     100/100           ▁▅▆█████   40 min at limit
 
         pending         1,910     ↑ 16×   ▁▁▂▃▅▅▇█   (normal ~120)
 
         start latency   p95 42s   ↑ 6×    ▁▁▂▄▆▆▇█   (normal ~7s)
 
-        worst queue     email-sends — 82% of pending
-
-        read: limit saturated → incoming work exceeds capacity → backlog grows
+        why:  82% of pending is email-sends
               runs are finishing at ~820/min
               nothing dead-lettered
 
-      EXECUTION   🟢 the runs that DO start are fine
+      🟢 EXECUTION   the runs that DO start are fine
 
-        failures 1.3% (normal ~1.1%) · durations normal
-        read: runs are completing normally
+      🟢 LIVENESS    fresh — telemetry current, updated 4s ago
 
-      LIVENESS    🟢 fresh — telemetry current, updated 4s ago
+        read: limit saturated → incoming work exceeds capacity → backlog grows
+              runs are completing normally
 
       → Raise the env concurrency limit
         Read concurrency docs
@@ -148,13 +144,22 @@ describe("health (Golden B — healthy)", () => {
     expect(renderReportMarkdown(vm)).toMatchInlineSnapshot(`
       "/report health        prod · last 1h · vs your 7d normal
 
-      🟢 Flow healthy  ·  🟢 Execution healthy  ·  🟢 data fresh
+      🟢 Flow healthy — starting normally
+
+        start latency   p95 6s      → flat   ██▅▅▂▁▁▂   (normal ~7s)
+
+        pending         84          → flat   ██▄▄▂▃▃▁   (normal ~120)
+
+        throughput      0/min
+          done          1,000/min
+          triggered     1,000/min
 
-      FLOW        🟢 starting normally — pending 84 (normal ~120) · starts p95 6s
+      🟢 EXECUTION   the runs that DO start are fine
 
-      EXECUTION   🟢 completing normally — failures 0.9% (normal ~1.1%) · durations normal
+      🟢 LIVENESS    fresh — telemetry current, updated 2s ago
 
-      LIVENESS    🟢 fresh — telemetry current, updated 2s ago
+        read: runs are starting on time
+              runs are completing normally
 
       → nothing to do"
     `);
@@ -176,11 +181,12 @@ describe("snapshot fallback path flags the estimated backlog trend", () => {
     },
   };
 
-  it("renders an (estimated) caveat on the proxy series, not a bare sparkline", () => {
+  it("renders an estimated caveat on the proxy series, not a bare sparkline", () => {
     const vm = interpret(snapshot);
     expect(vm.findings.find((f) => f.type === "flow")!.reason).toBe("backlog");
     const md = renderReportMarkdown(vm);
-    expect(md).toContain("(estimated)"); // the human surface signals the trend is a proxy
+    // the human surface signals the trend is a proxy
+    expect(md).toContain("(estimated from a proxy signal)");
   });
 });
 
@@ -509,7 +515,8 @@ describe("stale-telemetry trust guard covers flow (not just execution)", () => {
     // just repeat that), and no stale causal evidence (anomaly window) survives.
     const md = renderReportMarkdown(stale);
     expect(md).toContain("🔴 Flow unknown — data stale");
-    expect(md).toContain("🔴 flow can't be assessed");
+    expect(md).toContain("🔴 EXECUTION   execution can't be assessed");
+    expect(md).toContain("🚩 stale data"); // the report flags itself untrustworthy
     expect(md).not.toContain("(last 40 min)"); // anomaly window gone
   });
 
@@ -530,9 +537,9 @@ describe("stale-telemetry trust guard covers flow (not just execution)", () => {
 });
 
 describe("freshness unknown is distinct from lagging", () => {
-  it("renders 'data freshness unknown' in the summary, not 'data lagging'", () => {
+  it("renders the liveness section as 'freshness unknown', not 'data lagging'", () => {
     const md = renderReportMarkdown(interpret({ ...INPUT_A, liveness: { telemetryAgeMs: null } }));
-    expect(md).toContain("data freshness unknown");
+    expect(md).toContain("⚪ LIVENESS    freshness unknown");
     expect(md).not.toContain("data lagging");
   });