fix: drain pending cron deliveries after platform reconnect (#744)

- Added pending_deliveries.json buffer for failed cron deliveries
- drain_pending_deliveries() retries buffered messages on each tick
- Failed deliveries are buffered when adapter is down/reconnecting
- Entries expire after 3 retry attempts (prevents infinite buffering)
- Uses live adapter when available, falls back to standalone send

cron/scheduler.py

@@ -928,6 +928,14 @@ def tick(verbose: bool = True, adapters=None, loop=None) -> int:
         return 0
 
     try:
+        # Drain any pending deliveries from previous failed attempts
+        try:
+            drained = drain_pending_deliveries(adapters=adapters, loop=loop)
+            if drained:
+                logger.info("Drained %d pending delivery(ies)", drained)
+        except Exception as e:
+            logger.debug("Pending delivery drain error: %s", e)
+
         due_jobs = get_due_jobs()
 
         if verbose and not due_jobs:
@@ -969,6 +977,10 @@ def tick(verbose: bool = True, adapters=None, loop=None) -> int:
                         delivery_error = str(de)
                         logger.error("Delivery failed for job %s: %s", job["id"], de)
 
+                # Buffer failed deliveries for retry after reconnect
+                if delivery_error and should_deliver:
+                    _buffer_failed_delivery(job, deliver_content, delivery_error)
+
                 mark_job_run(job["id"], success, error, delivery_error=delivery_error)
                 executed += 1
 

gateway/run.py

@@ -1939,6 +1939,21 @@ class GatewayRunner:
                         )
                         logger.info("✓ %s reconnected successfully", platform.value)
 
+                        # Drain any pending cron deliveries that failed while disconnected
+                        try:
+                            from cron.scheduler import drain_pending_deliveries
+                            loop = asyncio.get_running_loop()
+                            drained = drain_pending_deliveries(
+                                adapters=self.adapters, loop=loop, max_retries=5,
+                            )
+                            if drained:
+                                logger.info(
+                                    "Drained %d pending cron delivery(ies) after %s reconnect",
+                                    drained, platform.value,
+                                )
+                        except Exception as e:
+                            logger.debug("Post-reconnect delivery drain error: %s", e)
+
                         # Rebuild channel directory with the new adapter
                         try:
                             from gateway.channel_directory import build_channel_directory