From ca54bd0b210e83298470d156f1e9b15586ff86dd Mon Sep 17 00:00:00 2001
From: Richard Kuo <rkuo@rkuo.com>
Date: Sun, 5 Jan 2025 20:02:05 -0800
Subject: [PATCH] add logging

---
 backend/onyx/background/celery/apps/primary.py      | 4 ++--
 backend/onyx/background/celery/tasks/vespa/tasks.py | 5 ++++-
 2 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/backend/onyx/background/celery/apps/primary.py b/backend/onyx/background/celery/apps/primary.py
index 62ea96658..caa697f88 100644
--- a/backend/onyx/background/celery/apps/primary.py
+++ b/backend/onyx/background/celery/apps/primary.py
@@ -89,12 +89,12 @@ def on_worker_init(sender: Any, **kwargs: Any) -> None:
     app_base.wait_for_db(sender, **kwargs)
     app_base.wait_for_vespa(sender, **kwargs)
 
+    logger.info("Running as the primary celery worker.")
+
     # Less startup checks in multi-tenant case
     if MULTI_TENANT:
         return
 
-    logger.info("Running as the primary celery worker.")
-
     # This is singleton work that should be done on startup exactly once
     # by the primary worker. This is unnecessary in the multi tenant scenario
     r = get_redis_client(tenant_id=None)
diff --git a/backend/onyx/background/celery/tasks/vespa/tasks.py b/backend/onyx/background/celery/tasks/vespa/tasks.py
index ff7b0b93f..c3c8c42ab 100644
--- a/backend/onyx/background/celery/tasks/vespa/tasks.py
+++ b/backend/onyx/background/celery/tasks/vespa/tasks.py
@@ -760,6 +760,8 @@ def monitor_vespa_sync(self: Task, tenant_id: str | None) -> bool:
 
     Returns True if the task actually did work, False if it exited early to prevent overlap
     """
+    task_logger.info(f"monitor_vespa_sync starting: tenant={tenant_id}")
+
     time_start = time.monotonic()
 
     timings: dict[str, float] = {}
@@ -775,6 +777,7 @@ def monitor_vespa_sync(self: Task, tenant_id: str | None) -> bool:
     try:
         # prevent overlapping tasks
         if not lock_beat.acquire(blocking=False):
+            task_logger.info("monitor_vespa_sync exiting due to overlap")
             return False
 
         # print current queue lengths
@@ -894,7 +897,7 @@ def monitor_vespa_sync(self: Task, tenant_id: str | None) -> bool:
             redis_lock_dump(lock_beat, r)
 
     time_elapsed = time.monotonic() - time_start
-    task_logger.debug(f"monitor_vespa_sync finished: elapsed={time_elapsed:.2f}")
+    task_logger.info(f"monitor_vespa_sync finished: elapsed={time_elapsed:.2f}")
     return True