Wavefront Django SDK

This SDK provides support for reporting out of the box metric, histograms and tracing from your Django based application. That data is reported to Wavefront via proxy or direct ingestion. That data will help you understand how your application is performing in production.

Install

pip install wavefront_django_sdk_python

Usage

Configure settings.py of your application to install Django SDK as follows:

# setting.py

from wavefront_pyformance.wavefront_reporter import WavefrontDirectReporter, WavefrontProxyReporter
from wavefront_sdk.common import ApplicationTags
from wavefront_django_sdk import DjangoTracing
from wavefront_opentracing_sdk import reporting, WavefrontTracer

INSTALLED_APPS = [
   '...',
   'wavefront_django_sdk',
   '...'
]

MIDDLEWARE = [
   'wavefront_django_sdk.middleware.WavefrontMiddleware',
   '...'
]

SOURCE = "{SOURCE}"

APPLICATION_TAGS = ApplicationTags(
   application="{APP_NAME}",
   service="{SERVICE_NAME}",
   cluster="{CLUSTER_NAME}",  # Optional
   shard="{SHARD_NAME}," , # Optional
   custom_tags={"location": "Oregon", "env": "Staging"}  # Optional
)

# Sending data via Direct Ingestion
WF_REPORTER = WavefrontDirectReporter(
   server="{ADDRESS}",
   token="{TOKEN}",
   reporting_interval=5,  # Optional, default value is 10 secs
   source=SOURCE,
   tags={"application": APPLICATION_TAGS.application}
).report_minute_distribution()

# Or, Sending data via Proxy
WF_REPORTER = WavefrontProxyReporter(
   host="{HOST}",
   port=2878,  # Optional, Wavefront Proxy running on 2878 by default
   reporting_interval=5,  # Optional, default value is 10 secs
   source=SOURCE,
   tags={"application": APPLICATION_TAGS.application}
).report_minute_distribution()

span_reporter = reporting.WavefrontSpanReporter(
   client=WF_REPORTER.wavefront_client,
   source=SOURCE,
)

OPENTRACING_TRACE_ALL = True  # Optional, default value is False

OPENTRACING_TRACING = DjangoTracing(WavefrontTracer(
   reporter=span_reporter, application_tags=APPLICATION_TAGS))

Out of the box metrics and histograms for your Django based application.

Assume you have the following API in your Django Application:

# urls.py
from django.urls import path
from . import views
 
urlpatterns = [
    path('style/<slug:id>/make', views.make_shirts, name="style/{id}/make")
]
 
# view.py
from django.http import HttpResponse
 
def make_shirts(request, id):
    return HttpResponse("completed", status=200)

Request Gauges

Entity Name	Entity Type	source	application	cluster	service	shard	django.resource.module	django.resource.func
django.request.style.id.make.GET.inflight.value	Gauge	host-1	Ordering	us-west-1	styling	primary	styling.views	make_shirts
django.total_requests.inflight.value	Gauge	host-1	Ordering	us-west-1	styling	primary	n/a	n/a

Granular Response related metrics

Entity Name	Entity Type	source	application	cluster	service	shard	django.resource.module	django.resource.func
django.response.style._id_.make.GET.200.cumulative.count	Counter	host-1	Ordering	us-west-1	styling	primary	styling.views	make_shirts
django.response.style._id_.make.GET.200.aggregated_per_shard.count	DeltaCounter	wavefront-provided	Ordering	us-west-1	styling	primary	styling.views	make_shirts
django.response.style._id_.make.GET.200.aggregated_per_service.count	DeltaCounter	wavefront-provided	Ordering	us-west-1	styling	n/a	styling.views	make_shirts
django.response.style._id_.make.GET.200.aggregated_per_cluster.count	DeltaCounter	wavefront-provided	Ordering	us-west-1	n/a	n/a	styling.views	make_shirts
django.response.style._id_.make.GET.200.aggregated_per_application.count	DeltaCounter	wavefront-provided	Ordering	n/a	n/a	n/a	styling.views	make_shirts

Granular Response related histograms

Entity Name	Entity Type	source	application	cluster	service	shard	django.resource.module	django.resource.func
django.response.style._id_.make.summary.GET.200.latency.m	WavefrontHistogram	host-1	Ordering	us-west-1	styling	primary	styling.views	make_shirts
django.response.style._id_.make.summary.GET.200.cpu_ns.m	WavefrontHistogram	host-1	Ordering	us-west-1	styling	primary	styling.views	make_shirts

Overall Response related metrics

This includes all the completed requests that returned a response (i.e. success + errors).

Entity Name	Entity Type	source	application	cluster	service	shard
django.response.completed.aggregated_per_source.count	Counter	host-1	Ordering	us-west-1	styling	primary
django.response.completed.aggregated_per_shard.count	DeltaCounter	wavefont-provided	Ordering	us-west-1	styling	primary
django.response.completed.aggregated_per_service.count	DeltaCounter	wavefont-provided	Ordering	us-west-1	styling	n/a
django.response.completed.aggregated_per_cluster.count	DeltaCounter	wavefont-provided	Ordering	us-west-1	n/a	n/a
django.response.completed.aggregated_per_application.count	DeltaCounter	wavefont-provided	Ordering	n/a	n/a	n/a

Overall Error Response related metrics

This includes all the completed requests that resulted in an error response (that is HTTP status code of 4xx or 5xx).

Entity Name	Entity Type	source	application	cluster	service	shard
django.response.errors.aggregated_per_source.count	Counter	host-1	Ordering	us-west-1	styling	primary
django.response.errors.aggregated_per_shard.count	DeltaCounter	wavefont-provided	Ordering	us-west-1	styling	primary
django.response.errors.aggregated_per_service.count	DeltaCounter	wavefont-provided	Ordering	us-west-1	styling	n/a
django.response.errors.aggregated_per_cluster.count	DeltaCounter	wavefont-provided	Ordering	us-west-1	n/a	n/a
django.response.errors.aggregated_per_application.count	DeltaCounter	wavefont-provided	Ordering	n/a	n/a	n/a

Tracing Spans

Every span will have the operation name as span name, start time in millisec along with duration in millisec. The following table includes all the rest attributes of generated tracing spans.

Span Tag Key	Span Tag Value
traceId	4a3dc181-d4ac-44bc-848b-133bb3811c31
parent	q908ddfe-4723-40a6-b1d3-1e85b60d9016
followsFrom	b768ddfe-4723-40a6-b1d3-1e85b60d9016
spanId	c908ddfe-4723-40a6-b1d3-1e85b60d9016
component	django
span.kind	server
application	Ordering
service	styling
cluster	us-west-1
shard	primary
location	Oregon (*custom tag)
env	Staging (*custom tag)
http.method	GET
http.url	http://{SERVER_ADDR}/style/{id}/make
http.status_code	502
error	True
django.resource.func	make_shirts
django.resource.module	styling.views

xshipeng/wavefront-django-sdk-python