Skip to main content

Logging formatter to tell Journald's streams the correct severity level

Project description

This Python module adds a prefix to log lines so that Journald knows which log level it comes from.

Features:

  • Can handle custom levels as well as Python's five standard ones.
  • Ensures that each line of a multiline log message is correctly prefixed.
  • Also applies a prefix to uncaught exceptions.
  • And prefixes any messages from the warnings module.

Usage in your Application

Install the latest stable version with

$ pip install priorityprefix

Or install the cutting-edge version with

$ pip install git+https://github.com/pscl4rke/priorityprefix.git

If you've already got a Formatter object for logging then you can wrap it with FormattingWrapper:

import priorityprefix
my_formatter = priorityprefix.FormattingWrapper(my_formatter)

For lightweight usage there is an install function which integrates nicely with basicConfig from the standard library

import logging
logging.basicConfig(level=logging.INFO)
import priorityprefix
priorityprefix.install()

By default install() will also override sys.excepthook so that any uncaught exceptions will get an error priority prefix added to their traceback before getting dumped to stderr.

Also, by default install() will also override warnings.formatwarning so that things like a DeprecationWarning will get a warning priority prefix before getting emitted.

Note that the logging library does not have a level that corresponds with systemd's NOTICE level. However you can create a custom level with a value of 25 to emit log records that are notices:

NOTICE = 25
my_logger.log(NOTICE, "Message goes here")

Usage with Journald

In short, you shouldn't need to change anything.

If setting up a Systemd service then you need to have the following settings in your service definition:

StandardOutput=journal
StandardError=journal
SyslogLevelPrefix=true

However these are the defaults, so you don't normally need to change anything.

Similarly, if you feed output through systemd-cat (which fulfils the same role as logger) then --level-prefix will also be enabled by default.

The Back-Story

To get your log messages into Systemd's Journald you have a number of options:

  • Send legacy Syslog datagrams to /dev/log and rely on Journald listening there and translating them.
  • Use something like libsystemd to send journal datagrams to /run/systemd/journal/socket.
  • Just print stuff to stdout or stderr. If Systemd has connected them up to the journal then each line will become a log record.

The last option is often the most popular because it's the easiest, but the major downside is that you can't specify any custom fields. You only have MESSAGE and the ones that Journald adds itself.

Except... you can precede lines with a special prefix to specify the severity of the message. This will be stripped off and turned into the PRIORITY.

Prefix PRIORITY
<7> Debug
<6> Informational
<5> Notice
<4> Warning
<3> Error
<2> Critical
<1> Alert
<0> Emergency

These prefices are inspired by Syslog, although it only encodes priority levels and does not include facility data.

The purpose of this library is to automatically add these prefices to every line of output.

The Roadmap

In many respects this project is now complete. Going forwards the things I want to achieve are:

  • Maintain compatibility with new versions of Python as they get released. Normally an annual thing.
  • Support automatic enabling or disabling of the prefix based on whether the JOURNAL_STREAM environment variable is present (supplied by Systemd) or absent (when the user is running things from a terminal). The complication of this is that the user would need to explicitly enable the prefix any time they were piping into systemd-cat.

Typechecking with Mypy et al

If you are checking the type annotations of your application you may get an error about priorityprefix being skipped because of missing stubs and markers. This can be suppressed in early versions of Mypy with a special comment:

import priorityprefix  # type: ignore

Some type annotations have been be added to this library. From Mypy version 1.14 onwards you can use --follow-untyped-imports or follow_untyped_imports = True to have them checked.

Unfortunately tools like Mypy adhere strictly to PEP 561 which requires the module to be fully refactored into a package before it will be recognised. Until a single-file module marker is available this library has no way to communicate to the tooling that type information is available.

Licence

This software library is released under the LGPL v3.0. It may be used by and distributed with both GPL and non-GPL applications.

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

priorityprefix-1.3.1.tar.gz (4.5 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

priorityprefix-1.3.1-py3-none-any.whl (4.5 kB view details)

Uploaded Python 3

File details

Details for the file priorityprefix-1.3.1.tar.gz.

File metadata

  • Download URL: priorityprefix-1.3.1.tar.gz
  • Upload date:
  • Size: 4.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.2

File hashes

Hashes for priorityprefix-1.3.1.tar.gz
Algorithm Hash digest
SHA256 3b71efdd408e39ff0afde52ce9d829f3245aec7d8a2c7c9a4326430920cd8caa
MD5 0fc03eadffdc7acdf88b235728f58a75
BLAKE2b-256 4f69e2ceab1555a923549be618e3bf1ac21ed365a5beb161115a4c3788c1c596

See more details on using hashes here.

File details

Details for the file priorityprefix-1.3.1-py3-none-any.whl.

File metadata

  • Download URL: priorityprefix-1.3.1-py3-none-any.whl
  • Upload date:
  • Size: 4.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.2

File hashes

Hashes for priorityprefix-1.3.1-py3-none-any.whl
Algorithm Hash digest
SHA256 b192535c98af6306303f20d360c2fd922ad067505089c6fc4d3ce406c2fd093a
MD5 a6a2f7b466d2727f424a6d503eb26cdf
BLAKE2b-256 baa881eaaabc89252177d4eed6b24cba186fddcc79e53b8a2b59074d80ac50cf

See more details on using hashes here.

Supported by

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