CODEWITHSUNDEEP
pythongoogle-cloud-platformgoogle-cloud-functionsgoogle-cloud-logging
C. Grd
C. Grd

Reputation: 68

Google Cloud Logging malfunctioning from cloud function

I am currently trying to deploy a cloud function triggered by Pub/Sub written in Python. Previously, we used loguru to log. I am now making the switch to the cloud logging. I thought it would be rather simple but am quite puzzled. Here is the code I deployed in a Cloud Function, just to try logging :

import base64
import logging
import google.cloud.logging as google_logging

def hello_pubsub(event, context):
     client = google_logging.Client()
     client.setup_logging()
     logging.debug("Starting function")
     logging.info("Hello")

     logging.warning("warning ! ")

     pubsub_message = base64.b64decode(event['data']).decode('utf-8')
     logging.info(pubsub_message)
     logging.error("Exit function")

I followed the documentation I could find on the subject (but the pages can show various methods, and are not very clear). Here is the result on the Logging interface :

Resulting logs in "global" tab

This is the result on the "Global" logs. Two questions here : why aren't the debug logs not shown, even if I explicitely set the log level as "debug" in the interface ? And why the logs are shown 1, 2 or 3 times, randomly ?

Now I try to display the logs for my Cloud Function only :

Resulting logs in "Cloud Function" logs for my function

This is getting worse, now the logs are displayed up to 5 times (and not even the same number of times than in the "Global" tab), the levels of informations are all wrong (logging.info results in 1 info line, 1 error line ; error and warning results in 2 errors lines...)

I imagine I must be doing something bad, but I can't see what, as what I am trying to do is fairly simple. Can somebody please help me ? Thanks !

EDIT : I did the mistake of putting the initialization of the client in the function, this explains that the logs were displayed more than once. One problem left is that the warnings are displayed as errors in the "Cloud Function" tabs, and displayed correctly in the "Global" tab. Do someone has an idea about this ?

Upvotes: 1

Views: 1043

Answers (2)

marian.vladoi
marian.vladoi

Reputation: 8074

You should use the Integration with Python logging module¶

import logging
import base64
import google.cloud.logging # Don't conflict with standard logging
from google.cloud.logging.handlers import CloudLoggingHandler
client = google.cloud.logging.Client()
handler = CloudLoggingHandler(client)
cloud_logger = logging.getLogger('cloudLogger')
cloud_logger.setLevel(logging.INFO) # defaults to WARN
cloud_logger.addHandler(handler)



def hello_pubsub(event, context):

     import logging
     cloud_logger.debug("Starting function")
     cloud_logger.info("Hello")

     cloud_logger.warning("warning ! ")

     pubsub_message = base64.b64decode(event['data']).decode('utf-8')
     cloud_logger.info(pubsub_message)
     cloud_logger.error("Exit function")

     return 'OK', 200

enter image description here

Upvotes: 1

Dustin Ingram
Dustin Ingram

Reputation: 21580

Try moving your setup outside of the function:

import base64
import logging
import google.cloud.logging as google_logging

client = google_logging.Client()
client.setup_logging()

def hello_pubsub(event, context):
     logging.debug("Starting function")
     logging.info("Hello")

     logging.warning("warning ! ")

     pubsub_message = base64.b64decode(event['data']).decode('utf-8')
     logging.info(pubsub_message)
     logging.error("Exit function")

As-is, you're adding new handlers on every request per instance.

Upvotes: 1

Related Questions