fix: use monitor-side checked_at timestamp instead of coordinator arrival time

apps/api/src/routes/pings.ts

@@ -65,13 +65,15 @@ export const ingest = new Elysia()
     const meta = body.meta ? { ...body.meta } : {};
     if (body.cert_expiry_days != null) meta.cert_expiry_days = body.cert_expiry_days;
 
+    const checkedAt = body.checked_at ? new Date(body.checked_at) : null;
     const scheduledAt = body.scheduled_at ? new Date(body.scheduled_at) : null;
     const jitterMs = body.jitter_ms ?? null;
 
     const [ping] = await sql`
-      INSERT INTO pings (monitor_id, scheduled_at, jitter_ms, status_code, latency_ms, up, error, meta, region, run_id)
+      INSERT INTO pings (monitor_id, checked_at, scheduled_at, jitter_ms, status_code, latency_ms, up, error, meta, region, run_id)
       VALUES (
         ${body.monitor_id},
+        ${checkedAt ?? sql`now()`},
         ${scheduledAt},
         ${jitterMs},
         ${body.status_code ?? null},
@@ -93,6 +95,7 @@ export const ingest = new Elysia()
   }, {
     body: t.Object({
       monitor_id:       t.String(),
+      checked_at:       t.Optional(t.Nullable(t.String())),
       scheduled_at:     t.Optional(t.Nullable(t.String())),
       jitter_ms:        t.Optional(t.Nullable(t.Number())),
       status_code:      t.Optional(t.Nullable(t.Number())),