fix: defer report scheduler checkpoint until enqueue succeeds

Move the REPORT_SCHEDULER_LAST_CHECK_CACHE_KEY checkpoint update from
before the enqueue loop to after each hour bucket is fully processed.

Previously the checkpoint was advanced to 'now' before any messages
were published. If the process crashed or the message queue was
unavailable, the pending hour buckets were permanently skipped because
the next run would start from the already-advanced checkpoint.

Changes:
- Advance checkpoint incrementally per hour bucket, only after all
  reports in that bucket have been successfully enqueued.
- Stop advancing checkpoint on first enqueue failure so the failed
  hour bucket (and any remaining ones) are retried on the next tick.
- Propagate cache.insert errors via '?' instead of silently ignoring
  them with 'let _ ='.
- When no hour boundaries need processing, still advance checkpoint
  to avoid re-scanning the same sub-hour window.

Fixes #1586

Author: ixchio

app-server/src/reports/scheduler.rs

@@ -96,12 +96,13 @@ async fn check_and_enqueue(
         now
     );
 
-    let _ = cache
-        .insert(REPORT_SCHEDULER_LAST_CHECK_CACHE_KEY, now.timestamp())
-        .await;
-
     let hours_to_check = hour_boundaries_between(last_check, now);
     if hours_to_check.is_empty() {
+        // No hour boundaries to process; advance checkpoint so the same
+        // sub-hour window is not re-scanned on the next tick.
+        cache
+            .insert(REPORT_SCHEDULER_LAST_CHECK_CACHE_KEY, now.timestamp())
+            .await?;
         return Ok(());
     }
     log::info!(
@@ -113,6 +114,7 @@ async fn check_and_enqueue(
     for (weekday, hour, triggered_at) in hours_to_check {
         let reports = get_reports_for_weekday_and_hour(pool, weekday, hour).await?;
 
+        let mut all_enqueued = true;
         for report in reports {
             let message = ReportTriggerMessage {
                 id: report.id,
@@ -129,8 +131,23 @@ async fn check_and_enqueue(
                     report.id,
                     e
                 );
+                all_enqueued = false;
             }
         }
+
+        if !all_enqueued {
+            // Stop advancing the checkpoint so this hour bucket (and any
+            // remaining ones) will be retried on the next tick.
+            return Err(anyhow::anyhow!(
+                "One or more reports failed to enqueue for hour {triggered_at}; checkpoint held for retry"
+            ));
+        }
+
+        // Advance checkpoint only after all reports in this hour bucket
+        // have been successfully enqueued.
+        cache
+            .insert(REPORT_SCHEDULER_LAST_CHECK_CACHE_KEY, triggered_at)
+            .await?;
     }
 
     Ok(())