i really hope this fixes stale

Author: transcental

nephthys/tasks/close_stale.py

@@ -12,38 +12,59 @@
 from prisma.enums import TicketStatus
 
 
-async def get_is_stale(ts: str) -> bool:
-    try:
-        replies = await env.slack_client.conversations_replies(
-            channel=env.slack_help_channel, ts=ts, limit=1000
-        )
-        last_reply = (
-            replies.get("messages", [])[-1] if replies.get("messages") else None
-        )
-        if not last_reply:
-            logging.error("No replies found - this should never happen")
-            await send_heartbeat(f"No replies found for ticket {ts}")
-            return False
-        return (
-            datetime.now(tz=timezone.utc)
-            - datetime.fromtimestamp(float(last_reply["ts"]), tz=timezone.utc)
-        ) > timedelta(days=3)
-    except SlackApiError as e:
-        if e.response["error"] == "ratelimited":
-            retry_after = int(e.response.headers.get("Retry-After", 1))
-            logging.warning(
-                f"Rate limited while fetching replies for ticket {ts}. Retrying after {retry_after} seconds."
-            )
-            await asyncio.sleep(retry_after)
-            return await get_is_stale(ts)
-        else:
-            logging.error(
-                f"Error fetching replies for ticket {ts}: {e.response['error']}"
+async def get_is_stale(ts: str, max_retries: int = 3) -> bool:
+    for attempt in range(max_retries):
+        try:
+            replies = await env.slack_client.conversations_replies(
+                channel=env.slack_help_channel, ts=ts, limit=1000
             )
-            await send_heartbeat(
-                f"Error fetching replies for ticket {ts}: {e.response['error']}"
+            last_reply = (
+                replies.get("messages", [])[-1] if replies.get("messages") else None
             )
-            return False
+            if not last_reply:
+                logging.error("No replies found - this should never happen")
+                await send_heartbeat(f"No replies found for ticket {ts}")
+                return False
+            return (
+                datetime.now(tz=timezone.utc)
+                - datetime.fromtimestamp(float(last_reply["ts"]), tz=timezone.utc)
+            ) > timedelta(days=3)
+        except SlackApiError as e:
+            if e.response["error"] == "ratelimited":
+                retry_after = int(e.response.headers.get("Retry-After", 1))
+                # Exponential backoff: wait longer on each retry
+                wait_time = retry_after * (2**attempt)
+                logging.warning(
+                    f"Rate limited while fetching replies for ticket {ts}. "
+                    f"Attempt {attempt + 1}/{max_retries}. Retrying after {wait_time} seconds."
+                )
+                await asyncio.sleep(wait_time)
+                if attempt == max_retries - 1:
+                    logging.error(f"Max retries exceeded for ticket {ts}")
+                    return False
+            if e.response["error"] == "thread_not_found":
+                logging.warning(
+                    f"Thread not found for ticket {ts}. This might be a deleted thread."
+                )
+                await send_heartbeat(f"Thread not found for ticket {ts}.")
+                await env.db.ticket.update(
+                    where={"msgTs": ts},
+                    data={
+                        "status": TicketStatus.CLOSED,
+                        "closedAt": datetime.now(),
+                        "closedBy": {"connect": {"slackId": env.slack_maintainer_id}},
+                    },
+                )
+                return False
+            else:
+                logging.error(
+                    f"Error fetching replies for ticket {ts}: {e.response['error']}"
+                )
+                await send_heartbeat(
+                    f"Error fetching replies for ticket {ts}: {e.response['error']}"
+                )
+                return False
+    return False
 
 
 async def close_stale_tickets():
@@ -58,21 +79,38 @@ async def close_stale_tickets():
     try:
         tickets = await env.db.ticket.find_many(
             where={"NOT": [{"status": TicketStatus.CLOSED}]},
-            include={
-                "openedBy": True,
-            },
+            include={"openedBy": True, "assignedTo": True},
         )
         stale = 0
 
-        for ticket in tickets:
-            if await get_is_stale(ticket.msgTs):
-                stale += 1
-                await resolve(
-                    ticket.msgTs,
-                    ticket.openedBy.slackId,  # type: ignore (this is valid - see include above)
-                    env.slack_client,
-                    stale=True,
-                )
+        # Process tickets in batches to avoid overwhelming the API
+        batch_size = 10
+        for i in range(0, len(tickets), batch_size):
+            batch = tickets[i : i + batch_size]
+            logging.info(
+                f"Processing batch {i // batch_size + 1}/{(len(tickets) + batch_size - 1) // batch_size}"
+            )
+
+            for ticket in batch:
+                await asyncio.sleep(1.2)  # Rate limiting delay
+
+                if await get_is_stale(ticket.msgTs):
+                    stale += 1
+                    resolver = (
+                        ticket.assignedToId
+                        if ticket.assignedToId
+                        else ticket.openedById
+                    )
+                    await resolve(
+                        ticket.msgTs,
+                        resolver,  # type: ignore (this is explicitly fetched in the db call)
+                        env.slack_client,
+                        stale=True,
+                    )
+
+            # Longer delay between batches
+            if i + batch_size < len(tickets):
+                await asyncio.sleep(5)
 
         await send_heartbeat(f"Closed {stale} stale tickets.")