CustomDimension not working

Describe your environment. Version: python3.6 Running as a container in RHEL VM. After importing opencensus.ext.azure.log_exporter docker image is built only when the Dockerfile is changed to have RUN yum install -y gcc rh-python36 rh-python36-python-devel.x86_64

Steps to reproduce. logger.py

import logging
import os
from opencensus.ext.azure.log_exporter import AzureLogHandler

_azhandler = AzureLogHandler(connection_string='InstrumentationKey={}'.format(os.getenv('app_insights_instrumentation_key')))
_azhandler.setLevel(logging.DEBUG)
_azhandler.setFormatter(logging.Formatter('%(levelname)s: %(message)s'))
_azlogger = logging.getLogger('FileUploader')
_azlogger.addHandler(_azhandler)

def info_customdimension(message, customDictionary):
    try:
        _azlogger.info(message, extra=customDictionary)
        print('Sent custom dictionary log')
    except Exception:
        print('Cannot send custom dictionary log')

Consumer of this logger file

 custom_dimensions = {}
 custom_dimensions['JobId'] = job_id
 custom_dimensions['FileUpload'] = True
 custom_properties = {}
 custom_properties['custom_dimensions'] = custom_dimensions

 info_customdimension("Successfully uploaded'), custom_properties)

What is the expected behavior? Application insights log should show custom dimensions properties along with the log message.

What is the actual behavior? Application insights log shows only log message.

Additional context. Add any other context about the problem here.

[raphaelauv]

@Izchen alright ! thank you.

I will open a ticket when I can from my admins … ( You don’t have permission to create a support request … )

[raphaelauv]

I’m having the same problem with custom_metric -> https://github.com/census-instrumentation/opencensus-python/pull/820

the dimensions do not show up in azure monitor.

code to reproduce :

from opencensus.ext.azure.common import utils
from opencensus.ext.azure.common.protocol import Data, Envelope, MetricData
from opencensus.ext.azure.log_exporter import AzureLogHandler


class AzureMetricPublisher(AzureLogHandler):

    def track_metric(self, metric_name, metric_value, namespace=None, dimensions=None):
        if namespace is None:
            namespace = metric_name
        if dimensions is None:
            dimensions = {}
        self._queue.put((metric_name, metric_value, namespace, dimensions), block=False)

    def log_record_to_envelope(self, record):
        if isinstance(record, tuple):
            envelope = Envelope(
                iKey=self.options.instrumentation_key,
                tags=dict(utils.azure_monitor_context),
                time=utils.to_iso_str(),
            )
            envelope.name = 'Microsoft.ApplicationInsights.Metric'
            name = record[0]
            value = record[1]
            ns = record[2]
            dimensions = record[3]
            data = MetricData(
                metrics=[
                    {'kind': None,
                     'name': name,
                     'ns': ns,
                     'max': value,
                     'count': 1,
                     'value': value,
                     'min': value,
                     'stdDev': 0}],
                properties=dimensions
            )
            print(data)
            print("properties :" + str(data.properties))
            envelope.data = Data(baseData=data, baseType='MetricData')
            return envelope

        return super().log_record_to_envelope(record)


metric_exporter = AzureMetricPublisher(connection_string="KEY")

metric_exporter.track_metric(metric_name='test_metric',
                             metric_value=42,
                             namespace="test_domain",
                             dimensions={"test_dimension": "value_dimension"})

If you run this code @lzchen do you see the dimension in azure monitor ?

Thank you