fix(crawler): queue chapter content in ascending number order (0.51.1)

Both enqueue paths now order by chapters.number so the cron tick and the
bookmark hook insert jobs from chapter 1 upward instead of source-discovery
or random-UUID order. The lease query tiebreaks on created_at so jobs
sharing a batch's scheduled_at come off the queue in insertion order,
propagating the enqueue intent through to dequeue. Concurrent workers
and per-CDN latency can still drift actual completion order.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

backend/tests/crawler_jobs.rs

@@ -531,6 +531,89 @@ async fn reap_done_deletes_old_rows_keeps_fresh(pool: PgPool) {
     assert_eq!(remaining, vec![fresh_id], "only fresh row remains");
 }
 
+#[sqlx::test(migrations = "./migrations")]
+async fn lease_ties_on_scheduled_at_break_by_created_at(pool: PgPool) {
+    // Locks in the tiebreaker that lets enqueue order survive the lease
+    // step: when many jobs share `scheduled_at` (the common cron-batch
+    // case), the worker must pick the earliest-inserted row, not whatever
+    // Postgres returns in heap order. The enqueue path inserts chapters
+    // in chapter-number order, so this tiebreaker is what makes "queue
+    // in rising order" observable at the dequeue side too.
+    let a = match jobs::enqueue(&pool, &chapter_content_payload(Uuid::new_v4()))
+        .await
+        .unwrap()
+    {
+        EnqueueResult::Inserted(id) => id,
+        _ => unreachable!(),
+    };
+    let b = match jobs::enqueue(&pool, &chapter_content_payload(Uuid::new_v4()))
+        .await
+        .unwrap()
+    {
+        EnqueueResult::Inserted(id) => id,
+        _ => unreachable!(),
+    };
+    let c = match jobs::enqueue(&pool, &chapter_content_payload(Uuid::new_v4()))
+        .await
+        .unwrap()
+    {
+        EnqueueResult::Inserted(id) => id,
+        _ => unreachable!(),
+    };
+
+    // Pin `scheduled_at` to a single literal instant (shared across all
+    // three rows — `now()` would yield a different microsecond per UPDATE
+    // and make scheduled_at the actual sort key). Reverse `created_at`
+    // against insertion order so heap order would give the wrong answer.
+    let shared_scheduled = chrono::Utc::now() - chrono::Duration::hours(1);
+    sqlx::query(
+        "UPDATE crawler_jobs \
+            SET scheduled_at = $2, \
+                created_at   = $3 \
+          WHERE id = $1",
+    )
+    .bind(a)
+    .bind(shared_scheduled)
+    .bind(chrono::Utc::now() - chrono::Duration::seconds(10))
+    .execute(&pool)
+    .await
+    .unwrap();
+    sqlx::query(
+        "UPDATE crawler_jobs \
+            SET scheduled_at = $2, \
+                created_at   = $3 \
+          WHERE id = $1",
+    )
+    .bind(b)
+    .bind(shared_scheduled)
+    .bind(chrono::Utc::now() - chrono::Duration::seconds(20))
+    .execute(&pool)
+    .await
+    .unwrap();
+    sqlx::query(
+        "UPDATE crawler_jobs \
+            SET scheduled_at = $2, \
+                created_at   = $3 \
+          WHERE id = $1",
+    )
+    .bind(c)
+    .bind(shared_scheduled)
+    .bind(chrono::Utc::now() - chrono::Duration::seconds(30))
+    .execute(&pool)
+    .await
+    .unwrap();
+
+    let leases = jobs::lease(&pool, None, 10, Duration::from_secs(60))
+        .await
+        .unwrap();
+    let order: Vec<Uuid> = leases.iter().map(|l| l.id).collect();
+    assert_eq!(
+        order,
+        vec![c, b, a],
+        "lease must return jobs in created_at order when scheduled_at ties"
+    );
+}
+
 #[sqlx::test(migrations = "./migrations")]
 async fn reap_done_zero_is_a_no_op(pool: PgPool) {
     let id = match jobs::enqueue(&pool, &chapter_content_payload(Uuid::new_v4()))