Skip to main content

Performance tracer middleware for Django

Project description

django-cavalry

A Performance Tracer!

CircleCI Codecov PyPI MIT License

Setup

  • Add 'cavalry.middleware.cavalry' to your middleware.
  • Add CAVALRY_ENABLED = bool(DEBUG) (or similar~) to your settings.

Settings – General

CAVALRY_ENABLED

(boolean, default False)

Master switch for the middleware.

This is toggleable during runtime. That is, it does not wholesale disable the middleware forever via the MiddlewareNotUsed mechanism.

CAVALRY_PROBABILITY

(float, default 1)

Probability (0..1) for a request to be traced; useful in production.

CAVALRY_DB_RECORD_STACKS

(boolean, default True)

Whether or not database stack traces should be recorded. Recording stack traces naturally has a performance impact.

Settings – Posting

The requests library must be available for posting to work.

You can install it by hand, or by using the [poster] extra while installing Cavalry.

CAVALRY_ELASTICSEARCH_URL_TEMPLATE

(string)

An URL template for posting payloads to Elasticsearch. {curly-braced} segments are interpolated using Python syntax. All fields in the payload are available, plus {ymd} is the YYYY-MM-DD of the current time. If falsy, no Elasticsearch posting is attempted.

An useful default might be 'http://localhost:9200/my-app-{ymd}/item'. This is easily ingestible by Kibana.

CAVALRY_THREADED_POST

(boolean, default False)

Whether or not to execute posting in another thread.

Enabling this has a positive performance impact in that formatting and submitting data to Elasticsearch data will not tie up request handling.

On the flipside, though, if the worker process dies before posting is complete, the trace is lost.

Also, if you're running on uWSGI, make sure enable-threads is set.

CAVALRY_POST_STACKS

(boolean, default True)

Whether or not post stack traces.

Not posting stack traces makes the ES payloads smaller.

Runtime

When

  • running in DEBUG, or
  • when you're a superuser, or
  • if request._cavalry_can_inject_stats is truthy (e.g. via a middleware of your own design)
    • Note that this may, especially in conjunction with _cavalry_stacks, reveal sensitive information. Be very diligent with that middleware.

Cavalry injects a small perf bar into each rendered HTML page, as well as a script segment that outputs SQL queries into the dev console.

By default, stack traces are not printed for the SQL queries; add the _cavalry_stacks query parameter to have them printed too.

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

django_cavalry-0.5.0.tar.gz (8.4 kB view details)

Uploaded Source

Built Distribution

django_cavalry-0.5.0-py3-none-any.whl (10.8 kB view details)

Uploaded Python 3

File details

Details for the file django_cavalry-0.5.0.tar.gz.

File metadata

  • Download URL: django_cavalry-0.5.0.tar.gz
  • Upload date:
  • Size: 8.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/5.1.1 CPython/3.12.6

File hashes

Hashes for django_cavalry-0.5.0.tar.gz
Algorithm Hash digest
SHA256 f064b9042c6e4858a4eda19151eafd8bd5bc821abf9e1a2b4d4573115c3b11ae
MD5 0aa2465103e0825025d4ffd630b8aa2a
BLAKE2b-256 eb13aafa4d506196572256e7a3e76271ad47f4e822a8292e4c60eddabfe5667d

See more details on using hashes here.

File details

Details for the file django_cavalry-0.5.0-py3-none-any.whl.

File metadata

File hashes

Hashes for django_cavalry-0.5.0-py3-none-any.whl
Algorithm Hash digest
SHA256 248d5e03b2181d46481cfded6903da3416db18ff1d5f746ebfd8f2809042e1d1
MD5 3b58429a5a4dc987645f81fa4bc54319
BLAKE2b-256 a697d9c5c3d981ba20a8e402e34987836011113d6aaaad1f4a188f0bc24e5b14

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page