tree: 856374f7c284e4f7c26f7ead1297c7dc8426ae43 [path history] [tgz]
  1. test/

Logging in infra.git


The infra_libs.logs package contains some code to simplify logging and make it consistent and easily configurable. Using it makes your code future-proof.

Offered features:

  • log level can be changed from the command-line.
  • too verbose modules can be blacklisted for easier debugging.
  • ensures a consistent log format.

A typical log line looks like:

[I2014-06-27T11:42:32.418716-07:00 7082 logs:71] this is the message

The first letter gives the severity of the message, followed by a timestamp with timezone information (iso8601), the process id, the current module name, and the thread id. After the closing square bracket comes the actual message.

Sample code

This is the standard way to set up logging so as to take advantage of the goodness provided by infra_libs.logs.

In top-level files (other example in

import argparse
import infra_libs.logs

parser = argparse.ArgumentParser()

options = parser.parse_args()

Logging messages should be done this way (other example in infra.libs.service_utils.outer_loop):

import logging
LOGGER = logging.getLogger(__name__)'great message')
LOGGER.error('terrible error')

Using logging.getLogger is a good practice in general (not restricted to using infra_libs.logs) because it allows for module blacklisting and other goodness. It should be done at import time. See also the official logging HOWTO. infra_libs.logs also formats the output of the root logger, but using this logger is not recommended.