Observe the behavior of your Tornado code
Project description
Helper classes for observing your Tornado-based HTTP code.
Source Code |
|
Download |
|
Documentation |
|
Issue Tracker |
Divák implements patterns for observing inter-service communication in distributed systems from both a client and a server perspective. The primary goal is to provide a nice framework for all aspects of observability:
measuring time spent during execution
counting interesting events
propagating tokens across inter-service communication that allow for tracing of messages
injecting tracing tokens into log messages
In the past I have approached the problem of observing data flows from the perspective of the individual aspects of logging, metrics recording, and error reporting. Divák takes a different approach. It provides the tools to include measurements in your code and the ability to propagate tokens that correlate all parts of a data flow across multiple services.
This particular library implements observation as an application level observer, a mix-in class that reads tokens from incoming HTTP requests, a wrapper that propagates tokens to other services when you make HTTP requests, and other similar helpers. The goal is to make it possible to easily instrument your application code in such a way that operational metrics are recorded, correlation identifiers are created and propagated between services, and the same identifiers are included in log messages. All of this using a single context-manager based instrumentation mechanism.
Here’s an example of the API that I am aiming for, The request handler uses two new mix-in classes: divak.api.RequestTracker and divak.api.Logger. The request tracker adds methods that implement operation tracing and the logger is essentially a logging.LoggingAdapter that injects the correlation identifiers into LogRecord instances. I also used the divak.api.HttpClientMixin which adds a method to send HTTP requests to other services that propagates the active span information.
The application also uses two mix-in classes: divak.api.Recorder and divak.api.ProbabilisticSampler. The record is responsible for submitting trace data out-of-band to whatever aggregators you add and the sampler is what determines how frequently to sample requests.
from tornado import gen, web
import divak.api
class MyHandler(divak.api.RequestTracker, divak.api.Logger,
divak.api.HttpClientMixin, web.RequestHandler):
def initialize(self):
super(self, MyHandler).initialize()
self.set_divak_name('foo')
@gen.coroutine
def get(self):
with self.record_operation('query-db'):
rows = yield self.query('...')
self.add_divak_tag('query-db.rowcount', len(rows))
output = []
for row in rows:
self.logger.debug('requesting %r', row['id'])
response = yield self.send_request(
'http://.../?id={}'.format(row['id']),
method='GET')
output.append(response.body)
self.set_status(200)
self.send_response(output)
class Application(divak.api.Recorder, divak.api.ProbabilisticSampler,
web.Application):
def __init__(self, *args, **kwargs):
handlers = [web.url(r'/', MyHandler)]
super(Application, self).__init__(handlers, *args, **kwargs)
self.set_divak_service('my-service')
self.set_divak_sample_probability(1.0 / 1000.0)
self.add_divak_reporter(
divak.api.ZipkinReporter('http://127.0.0.1:9411/api/v2'))
self.add_divak_propagator(divak.api.ZipkinPropagation())
This example will result in zipkin tracing for GET / requests which record spans for the database query and each HTTP API call.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file divak-tornado-0.0.3.tar.gz
.
File metadata
- Download URL: divak-tornado-0.0.3.tar.gz
- Upload date:
- Size: 17.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 337cb37d236169124dc4e1659df4ce1e00fa465445b02a0bc2925585bbbca93c |
|
MD5 | 28a1f16c9480f4234d0e7e3f1f8c44d8 |
|
BLAKE2b-256 | cd57fe575414fd07b78e602333f2603a2a7b7c9059222e40ba2e078189ba9d8b |
File details
Details for the file divak_tornado-0.0.3-py2.py3-none-any.whl
.
File metadata
- Download URL: divak_tornado-0.0.3-py2.py3-none-any.whl
- Upload date:
- Size: 11.9 kB
- Tags: Python 2, Python 3
- Uploaded using Trusted Publishing? No
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | b2de49c2a16581ad7a664e6304e5fb3c44dd25efc5b88f04d292e2e5cc35a247 |
|
MD5 | e19afd37b9c2b5d690050b1a60b67907 |
|
BLAKE2b-256 | 0c2f167ac8663bc21976e6cde073947743765f236723f0eb5e20c874d62b74ed |