Actually, it looks like it was because I registered the same logging handler twice, which brings me to another question: for azure functions, since the functions are hosted on the same vm, will logger = logging.getLogger(name) work? i.e. since it is on the same vm, will I have to do something like: logger = logging.getLogger(f'{name}-{uuid()}') to ensure that each instance of the logger doesn't get registered multiple handlers?
i dont think name has anything to do with the vm
think of the tagmap as a "pool" of key-values that can be selected from
For each "view" that you create, you can specify the keys t hat you want for that specific metric
The corresponding metric telemetry will search the tagmap for that key and find that value, and then that keyvalue pair would be one of the custom dimensions the metric telemetry will have
So in the carrots example in the read me, if I wanted to add an customDimension called “environment”, I would do the following:
tag_key = tag_key.TagKey("environment")
tag_map = tag_map_module.TagMap().insert(tag_key, tag_value.TagValue(“staging”))Then on calling the .record() method for the measure
mmap.record(tag_map)This is what I have done and the code works I just dont see the custom dimension in app insights. Am I going about this the right way?
CARROTS_VIEW = view_module.View("carrots_view",
"number of carrots",
["environment"],
CARROTS_MEASURE,
aggregation_module.CountAggregation())
"number of carrots",
["environment"],
CARROTS_MEASURE,
aggregation_module.CountAggregation())
Hi I am trying to use opencensus-ext-azure in the following way:
import logging
import logging.handlers
from opencensus.ext.azure.log_exporter import AzureLogHandler
log = logging.getLogger(__name__)
root_logger = logging.getLogger()
telemetry_handler = AzureLogHandler(
connection_string='InstrumentationKey=00000-000-000-000-00000')
telemetry_handler.setLevel(logging.INFO)
root_logger.addHandler(telemetry_handler)
def close_everything():
log.info('closing everything')
try:
some_error
except:
log.exception('some exception', exc_info=False)
pass
import atexit
atexit.register(close_everything)
atexit.register(logging.shutdown)The issue is that python session hangs at exit(). I think line which is creating problem is log.exception('some exception', exc_info=False). Can you please help. Thanks !
It also seems like
close_everything isn't being called correct?
Oh on exit I see nevermind LD
As well, the
AzureLogHandler registers a flush operation upon exit of the main script, how long is the session hanging for?
@lakhotiaharshit can you file an issue on the OC github with details of your problem? https://github.com/census-instrumentation/opencensus-python/issues
@lzchen I made the issue census-instrumentation/opencensus-python#881 . Also just to give you a bit more idea on our exact use case, here is the QCoDeS/Qcodes#1968 where we are migrating from applicationinsights to opencensus. Thanks a lot for your help in advance.
@lzchen Things got fixed.
1 ) logging.shutdown is called (this was the last thing we registered and atexit follows opposite order, which I was not aware of. I was thinking first close_everthing happens and logging.shutdown)
2) Then close_everything is called
3) Then logging.shutdown is called again (since this is registered as soon as logging is imported)
try removing logging.shutdown your suggestion worked. In the background this was happening:1 ) logging.shutdown is called (this was the last thing we registered and atexit follows opposite order, which I was not aware of. I was thinking first close_everthing happens and logging.shutdown)
2) Then close_everything is called
3) Then logging.shutdown is called again (since this is registered as soon as logging is imported)