I have an issue with k8s startup probes where when it meets mongoClient (pymongo), my pods will stuck in unready state

def test_mongodb(conn_str):      logging.debug("testing Mongo...")        try:                   start_time = time.time()  # Start timing                   client = MongoClient(conn_str)                   info = client.server_info()                   end_time = time.time()  # End timing                   elapsed_time = end_time - start_time                   logging.debug(f"MongoDB connection check took {elapsed_time:.2f} seconds")        logging.debug("Mongodb server info: %s", info)                   print("MongoDB Server Information:", info)                     return True          except Exception as e:                            logging.debug(f"mongo failed: {str(e)}")                      return Falsetest_mongodb(conn_str)

Helm chart for the startup probe

startupProbe:  enabled: true  initialDelaySeconds: 30  periodSeconds: 60  failureThreshold: 6  exec:    command:    - python    - /app/start.py

my pods will become unready state. so its not even an uncaught error. there is no error to begin with. I can see on the logs that the probe return success.

the helm release logs would give me: "context deadline error"/"client rate limiter".

but if i comment the line:

# client = MongoClient(conn_str) # info = client.server_info()

my pods will work just fine with no restart by the probes.