OpenCensus for Python. OpenCensus provides a framework to measure a server's resource usage and collect performance stats. This repository contains Python related utilities and supporting software needed by OpenCensus.
Install the opencensus package using pip or pipenv:
pip install opencensus pipenv install opencensus
Initialize a tracer for your application:
from opencensus.trace.tracer import Tracer from opencensus.trace.samplers import AlwaysOnSampler tracer = Tracer(sampler=AlwaysOnSampler())
Initialize a view_manager and a stats_recorder for your application:
from opencensus.stats import stats as stats_module stats = stats_module.stats view_manager = stats.view_manager stats_recorder = stats.stats_recorder
You can collect traces using the Tracer
context manager:
from opencensus.trace.tracer import Tracer
from opencensus.trace.samplers import AlwaysOnSampler
# Initialize a tracer, by default using the `PrintExporter`
tracer = Tracer(sampler=AlwaysOnSampler())
# Example for creating nested spans
with tracer.span(name='span1') as span1:
do_something_to_trace()
with span1.span(name='span1_child1') as span1_child1:
do_something_to_trace()
with span1.span(name='span1_child2') as span1_child2:
do_something_to_trace()
with tracer.span(name='span2') as span2:
do_something_to_trace()
OpenCensus will collect everything within the with
statement as a single span.
Alternatively, you can explicitly start and end a span:
from opencensus.trace.tracer import Tracer
from opencensus.trace.samplers import AlwaysOnSampler
# Initialize a tracer, by default using the `PrintExporter`
tracer = Tracer(sampler=AlwaysOnSampler())
tracer.start_span(name='span1')
do_something_to_trace()
tracer.end_span()
There are several things you can customize in OpenCensus:
- Blacklist, which excludes certain hosts and paths from being tracked. By default, the health check path for the App Engine flexible environment is not tracked, you can turn it on by excluding it from the blacklist setting.
- Exporter, which sends the traces.
By default, the traces are printed to stdout in JSON format. You can choose
different exporters to send the traces to. There are three built-in exporters,
which are
PrintExporter
,FileExporter
andLoggingExporter
, the other exporters are provided as extensions. - Sampler, which determines how traces are sampled.
The default sampler is the
ProbabilitySampler
, which samples (i.e. enables tracing for) a percentage of all requests. Sampling is deterministic according to the trace ID. To force sampling for all requests, or to prevent any request from being sampled, seeAlwaysOnSampler
andAlwaysOffSampler
. - Propagator, which serializes and deserializes the
SpanContext
and its headers. The default propagator isTraceContextPropagator
, other propagators includeBinaryFormatPropagator
,GoogleCloudFormatPropagator
andTextFormatPropagator
.
You can customize while initializing a tracer.
import requests
from opencensus.trace import config_integration
from opencensus.trace import file_exporter
from opencensus.trace import tracer as tracer_module
from opencensus.trace.propagation import google_cloud_format
from opencensus.trace.samplers import ProbabilitySampler
config_integration.trace_integrations(['httplib'])
tracer = tracer_module.Tracer(
exporter=file_exporter.FileExporter(file_name='traces'),
propagator=google_cloud_format.GoogleCloudFormatPropagator(),
sampler=ProbabilitySampler(rate=0.5),
)
with tracer.span(name='parent'):
with tracer.span(name='child'):
response = requests.get('http://localhost:5000')
You can use a configuration file for Flask/Django/Pyramid. For more information, please read the individual integration documentation.
'OPENCENSUS': {
'TRACE': {
'BLACKLIST_HOSTNAMES': ['localhost', '127.0.0.1'],
'BLACKLIST_PATHS': ['_ah/health'],
'SAMPLER': 'opencensus.trace.samplers.ProbabilitySampler(rate=1)',
'EXPORTER': '''opencensus.ext.ocagent.trace_exporter.TraceExporter(
service_name='foobar',
)''',
'PROPAGATOR': 'opencensus.trace.propagation.google_cloud_format.GoogleCloudFormatPropagator()',
}
}
OpenCensus supports integration with popular web frameworks, client libraries and built-in libraries.
- Django
- Flask
- gevent
- Google Cloud Client Libraries
- gRPC
- httplib
- logging
- MySQL
- PostgreSQL
- pymongo
- PyMySQL
- Pyramid
- requests
- SQLAlchemy
- threading
This library follows Semantic Versioning.
GA: Libraries defined at a GA quality level are stable, and will not introduce backwards-incompatible changes in any minor or patch releases. We will address issues and requests with the highest priority. If we were to make a backwards-incompatible changes on an API, we will first mark the existing API as deprecated and keep it for 18 months before removing it.
Beta: Libraries defined at a Beta quality level are expected to be mostly stable and we're working towards their release candidate. We will address issues and requests with a higher priority. There may be backwards incompatible changes in a minor version release, though not in a patch release. If an element is part of an API that is only meant to be used by exporters or other opencensus libraries, then there is no deprecation period. Otherwise, we will deprecate it for 18 months before removing it, if possible.