opencensus-ext-azure 1.1.15

Correlation

You can enrich the logs with trace IDs and span IDs by using the logging integration.

• Install the logging integration package using pip install opencensus-ext-logging.

importloggingfromopencensus.ext.azure.log_exporterimport AzureLogHandlerfromopencensus.ext.azure.trace_exporterimport AzureExporterfromopencensus.traceimport config_integrationfromopencensus.trace.samplersimport ProbabilitySamplerfromopencensus.trace.tracerimport Tracerconfig_integration.trace_integrations(['logging'])logger = logging.getLogger(__name__)handler = AzureLogHandler(connection_string='InstrumentationKey=<your-instrumentation_key-here>')handler.setFormatter(logging.Formatter('%(traceId)s%(spanId)s%(message)s'))logger.addHandler(handler)tracer = Tracer( exporter=AzureExporter(connection_string='InstrumentationKey=<your-instrumentation_key-here>'), sampler=ProbabilitySampler(1.0))logger.warning('Before the span')with tracer.span(name='test'): logger.warning('In the span')logger.warning('After the span')

Custom Properties

You can also add custom properties to your log messages in the extra keyword argument using the custom_dimensions field.

WARNING: For this feature to work, you need to pass a dictionary to the custom_dimensions field. If you pass arguments of any other type, the logger will ignore them.

importloggingfromopencensus.ext.azure.log_exporterimport AzureLogHandlerlogger = logging.getLogger(__name__)logger.addHandler(AzureLogHandler(connection_string='InstrumentationKey=<your-instrumentation_key-here>'))properties = {'custom_dimensions': {'key_1': 'value_1', 'key_2': 'value_2'}}logger.warning('action', extra=properties)

Modifying Logs

• You can pass a callback function to the exporter to process telemetry before it is exported.

• Your callback function can return False if you do not want this envelope exported.

• Your callback function must accept an envelope data type as its parameter.

• You can see the schema for Azure Monitor data types in the envelopes here.

• The AzureLogHandler handles ExceptionData and MessageData data types.

importloggingfromopencensus.ext.azure.log_exporterimport AzureLogHandlerlogger = logging.getLogger(__name__)# Callback function to append '_hello' to each log message telemetrydefcallback_function(envelope): envelope.data.baseData.message += '_hello' return Truehandler = AzureLogHandler(connection_string='InstrumentationKey=<your-instrumentation_key-here>')handler.add_telemetry_processor(callback_function)logger.addHandler(handler)logger.warning('Hello, World!')

Events

You can send customEvent telemetry in exactly the same way you would send trace telemetry except using the AzureEventHandler instead.

importloggingfromopencensus.ext.azure.log_exporterimport AzureEventHandlerlogger = logging.getLogger(__name__)logger.addHandler(AzureEventHandler(connection_string='InstrumentationKey=<your-instrumentation_key-here>'))logger.setLevel(logging.INFO)logger.info('Hello, World!')

Performance counters

The exporter also includes a set of performance counters that are exported to Azure Monitor by default.

importpsutilimporttimefromopencensus.ext.azureimport metrics_exporterdefmain(): # Performance counters are sent by default. You can disable performance counters by # passing in enable_standard_metrics=False into the constructor of # new_metrics_exporter() _exporter = metrics_exporter.new_metrics_exporter( connection_string='InstrumentationKey=<your-instrumentation-key-here>', export_interval=60, ) for i in range(100): print(psutil.virtual_memory()) time.sleep(5) print("Done recording metrics")if __name__ == "__main__": main()

Below is a list of performance counters that are currently available:

• Available Memory (bytes)

• CPU Processor Time (percentage)

• Incoming Request Rate (per second)

• Incoming Request Average Execution Time (milliseconds)

• Process CPU Usage (percentage)

• Process Private Bytes (bytes)

Modifying Metrics

• You can pass a callback function to the exporter to process telemetry before it is exported.

• Your callback function can return False if you do not want this envelope exported.

• Your callback function must accept an envelope data type as its parameter.

• You can see the schema for Azure Monitor data types in the envelopes here.

• The MetricsExporter handles MetricData data types.

fromopencensus.ext.azureimport metrics_exporterfromopencensus.statsimport aggregation as aggregation_modulefromopencensus.statsimport measure as measure_modulefromopencensus.statsimport stats as stats_modulefromopencensus.statsimport view as view_modulefromopencensus.tagsimport tag_map as tag_map_modulestats = stats_module.statsview_manager = stats.view_managerstats_recorder = stats.stats_recorderCARROTS_MEASURE = measure_module.MeasureInt("carrots", "number of carrots", "carrots")CARROTS_VIEW = view_module.View("carrots_view", "number of carrots", [], CARROTS_MEASURE, aggregation_module.CountAggregation())# Callback function to only export the metric if value is greater than 0defcallback_function(envelope): return envelope.data.baseData.metrics[0].value > 0defmain(): # Enable metrics # Set the interval in seconds to 60s, which is the time interval application insights # aggregates your metrics exporter = metrics_exporter.new_metrics_exporter( connection_string='InstrumentationKey=<your-instrumentation-key-here>', export_interval=60, ) exporter.add_telemetry_processor(callback_function) view_manager.register_exporter(exporter) view_manager.register_view(CARROTS_VIEW) mmap = stats_recorder.new_measurement_map() tmap = tag_map_module.TagMap() mmap.measure_int_put(CARROTS_MEASURE, 1000) mmap.record(tmap) print("Done recording metrics")if __name__ == "__main__": main()

Integrations

OpenCensus also supports several integrations which allows OpenCensus to integrate with third party libraries.

This example shows how to integrate with the requests library.

• Install the requests integration package using pip install opencensus-ext-requests.

importrequestsfromopencensus.ext.azure.trace_exporterimport AzureExporterfromopencensus.traceimport config_integrationfromopencensus.trace.samplersimport ProbabilitySamplerfromopencensus.trace.tracerimport Tracerconfig_integration.trace_integrations(['requests'])tracer = Tracer( exporter=AzureExporter( connection_string='InstrumentationKey=<your-instrumentation-key-here>', ), sampler=ProbabilitySampler(1.0),)with tracer.span(name='parent'): response = requests.get(url='https://www.wikipedia.org/wiki/Rabbit')

Modifying Traces

• You can pass a callback function to the exporter to process telemetry before it is exported.

• Your callback function can return False if you do not want this envelope exported.

• Your callback function must accept an envelope data type as its parameter.

• You can see the schema for Azure Monitor data types in the envelopes here.

• The AzureExporter handles Data data types.

importrequestsfromopencensus.ext.azure.trace_exporterimport AzureExporterfromopencensus.traceimport config_integrationfromopencensus.trace.samplersimport ProbabilitySamplerfromopencensus.trace.tracerimport Tracerconfig_integration.trace_integrations(['requests'])# Callback function to add os_type: linux to span propertiesdefcallback_function(envelope): envelope.data.baseData.properties['os_type'] = 'linux' return Trueexporter = AzureExporter( connection_string='InstrumentationKey=<your-instrumentation-key-here>')exporter.add_telemetry_processor(callback_function)tracer = Tracer(exporter=exporter, sampler=ProbabilitySampler(1.0))with tracer.span(name='parent'): response = requests.get(url='https://www.wikipedia.org/wiki/Rabbit')

Integrate with Azure Functions

Users who want to capture custom telemetry in Azure Functions environments are encouraged to used the OpenCensus Python Azure Functions extension. More details can be found in this document.