A startup script has been provided to demonstrated how to properly provision a GCE

instance with fluentd configured. Note the intallation of fluentd, the addition of the config file,

and the restarting of the fluetnd service. You can start an instance using

it like this:

gcloud compute instances create example-instance --metadata-from-file startup-script=startup_script.sh

or simply use it as reference when creating your own instance.

After fluentd is configured, main.py could be used to simulate an error:

gcloud compute copy-files main.py example-instance:~/main.py

Then,

gcloud compute ssh example-instance

python ~/main.py

And you will see the message in the Errors Console.

import traceback

import fluent.event

import fluent.sender

fluent.sender.setup('myapp', host='localhost', port=24224)

def report(ex):

data = {}

data['message'] = '{0}'.format(ex)

data['serviceContext'] = {'service': 'myapp'}

# ... add more metadata

fluent.event.Event('errors', data)

# report exception data using:

report(traceback.format_exc())

def test_error_sends(event_mock):

event_mock.Event.assert_called_once_with(mock.ANY, mock.ANY)