skip to navigation
skip to content

Not Logged In

mod_wsgi-metrics 1.1.1

Metrics package for Apache/mod_wsgi.

Note that this package is no longer being maintained at this point. At some point it will be done over to use a different mechanism for visualising data so as not to depend on a separate service.

The mod_wsgi-metrics package is an add on package for Apache/mod_wsgi. It generates metric information about the run time performance of Apache and mod_wsgi. At least mod_wsgi version 4.2.0 is required.

In this version, metrics collected cover the performance of the Apache web server as a whole. In future versions additional metrics will be added which monitor aspects of mod_wsgi itself.

At the present time the package provides a plugin for the New Relic Platform. This plugin is distinct from New Relic’s own Python agent for use in monitoring Python web applications. The plugin instead focuses on metrics specific to Apache and mod_wsgi. The information from these metrics can be used to help in tuning your Apache/mod_wsgi installation for best performance.

The New Relic Platform is a free feature of New Relic and so in order to use this plugin for Apache/mod_wsgi, you do not need to have a paid account for New Relic.

Using the plugin with a mod_wsgi express installation

When using mod_wsgi express, the plugin will be automatically started and will report data when using the builtin support of mod_wsgi express for New Relic. See the mod_wsgi express documentation for more information on starting it with New Relic support enabled.

Using the plugin with a standard mod_wsgi installation

If you have installed mod_wsgi as an Apache module direct into your Apache installation, or have installed an operating system binary package, and are configuring Apache manually to host your Python web application, additional setup will be required to enable the plugin.

The steps for manually enabling the plugin are as follows:

1. Create a Python script file called server-metrics.py. In that file place:

import logging

logging.basicConfig(level=logging.INFO,
    format='%(name)s (pid=%(process)d, level=%(levelname)s): %(message)s')

from mod_wsgi.metrics.newrelic import Agent

config_file = '/some/path/newrelic.ini'

agent = Agent(config_file=config_file)
agent.start()

This would normally be placed along side your Python web application code.

The config_file variable should be set to the location of the newrelic.ini agent configuration file you created for use with the New Relic Python agent.

Alternatively, you can set the New Relic license key and application name to report to within the Python script file:

license_key = 'YOUR-NEW-RELIC-LICENSE-KEY'
app_name = 'THE-APPLICATION-NAME-TO-REPORT-AGAINST'

agent = Agent(app_name=app_name, license_key=license_key)
agent.start()

This Python script file would normally be placed along side your Python web application code.

2. Ensure that the mod_status module is being loaded into Apache and that ExtendedStatus is On:

LoadModule status_module modules/mod_status.so
ExtendedStatus On

The exact way in which this needs to be done will differ between Apache installations, especially with Apache installations provided by a Linux distribution. You should therefore look closely at how this is managed for your Apache installation.

Note that it is only necessary to load mod_status and enable ExtendedStatus. It is not necessary to expose the traditional /server-status URL generally associated with the use of mod_status as the plugin will not use that. Instead the plugin obtains the information from the mod_wsgi module. The mod_status module still has to be loaded though, otherwise Apache will not collect the information that is required.

3. Create a dedicated mod_wsgi daemon process group using the WSGIDaemonProcess directive. This should have only a single process and a single thread. It should also enable visibility of internal server metrics from mod_wsgi using the server-metrics option:

WSGIDaemonProcess newrelic display-name=%{GROUP} \
    processes=1 threads=1 server-metrics=On

This daemon process group should not be used to host your actual Python web application.

4. Specify that the server-metrics.py Python script file you created should be loaded when Apache is (re)started using the WSGIImportScript directive:

WSGIImportScript /some/path/server-metrics.py \
    process-group=newrelic application-group=%{GLOBAL}

The path should match where you saved the server-metrics.py script. The process-group option should match the name of the daemon process group created with using the WSGIDaemonProcess directive.

4. Restart Apache. Within the New Relic UI you should automatically see a new entry appear in the left hand navigation bar labelled ‘mod_wsgi’. The reported data will then appear under the application name used.

 
File Type Py Version Uploaded on Size
mod_wsgi-metrics-1.1.1.tar.gz (md5) Source 2014-12-24 17KB
  • Downloads (All Versions):
  • 14 downloads in the last day
  • 156 downloads in the last week
  • 521 downloads in the last month