Structured JSON logging
Project description
Sift Log - JSON logging adapter for Python (now in color)
Features
- Tag log statements with arbitrary values for easier grouping and analysis
- Add keyword arguments that are converted to JSON values
- Variable substitution
- Specifies where log calls are made from
- Meant to be used with core Python logging (formatters, handlers, etc)
- Colorized logs on a console (POSIX only)
TRACE
log level built-in
Examples
A simple log message
log.info('Hello')
{"msg": "Hello", "time": "12-12-14 10:12:01 EST", "level": "INFO", "loc": "test:log_test:20"}
Logging with tags
log.debug('Creating new user', 'MONGO', 'STORAGE')
{"msg": "Creating new user", "time": "12-12-14 10:12:09 EST", "tags": ["MONGO", "STORAGE"], "level": "DEBUG", "loc": "test:log_test:20"}
Appending more data
log.debug('Some key', is_admin=True, username='papito')
{"msg": "Some key", "is_admin": true, "username": "papito", "time": "12-12-14 10:12:04 EST", "level": "DEBUG", "loc": "test:log_test:20"}
String substitution
log.debug('User "$username" admin? $is_admin', is_admin=False, username='fez')
{"msg": "User \"fez\" admin? False", "username": "fez", "is_admin": false, "time": "12-12-14 10:12:18 EST", "level": "DEBUG", "loc": "test:log_test:20"}
Setup
Logging to console
import sys
import logging
from siftlog import SiftLog
logger = logging.getLogger()
logger.setLevel(logging.INFO)
handler = logging.StreamHandler(sys.stdout)
logger.addHandler(handler)
log = SiftLog(logger)
In this fashion, you can direct the JSON logs to any logging handler
Color
For enhanced flamboyancy, attach the ColorStreamHandler
to your logger. The output will not have color if the logs
are being output to a file, or on systems that are not POSIX (will not work on Windows for now).
from siftlog import SiftLog, ColorStreamHandler
logger = logging.getLogger()
handler = ColorStreamHandler(sys.stdout)
logger.addHandler(handler)
log = SiftLog(logger)
Performance
While the above should play, it's highly recommended that the color handler is only attached conditionally for local development. Too many log statements could otherwise become expensive in terms of CPU.
Different colors
You can change font background, text color, and boldness:
from siftlog import ColorStreamHandler
handler = ColorStreamHandler(sys.stdout)
handler.set_color(
logging.DEBUG, bg=handler.WHITE, fg=handler.BLUE, bold=True
)
Supported colors
- ColorStreamHandler.BLACK
- ColorStreamHandler.RED
- ColorStreamHandler.GREEN
- ColorStreamHandler.YELLOW
- ColorStreamHandler.BLUE
- ColorStreamHandler.MAGENTA
- ColorStreamHandler.CYAN
- ColorStreamHandler.WHITE
Constants (re-occuring values)
You can define constants that will appear in every single log message. This is useful, for example, if you'd like to log process PID and hostname with every log message (recommended). This is done upon log adapter initialization:
import os
from siftlog import SiftLog
log = SiftLog(logger, pid=os.getpid(), env='INTEGRATION')
{"msg": "And here I am", "time": "12-12-14 11:12:24 EST", "pid": 37463, "env": "INTEGRATION", "level": "INFO"}
Custom time format
log = SiftLog(logger)
SiftLog.TIME_FORMAT = '%d-%m-%y %H:%m:%S %Z'
Define the format as accepted by time.strftime()
Custom location format
log = SiftLog(logger)
SiftLog.LOCATION_FORMAT = '$module:$method:$line_no'
The format should be a string containing any of the following variables:
$file
$line_no
$method
$module
Custom core key names
Core keys, such as msg
and level
can be overridden, if they clash with common keys you might be using.
The following can be redefined:
- SiftLog.MESSAGE (default
msg
) - SiftLog.LEVEL (default
level
) - SiftLog.LOCATION (default
loc
) - SiftLog.TAGS (default
tags
) - SiftLog.TIME (default
time
)
As in:
log = SiftLog(logger)
SiftLog.log.MESSAGE = "MESSAGE"
Development flow
Poetry
is used to manage the dependencies.
Most things can be accessed via the Makefile, if you have Make installed.
# use the right Python
poetry use path/to/python/3.8-ish
# make sure correct Python is used
make info
# install dependencies
make install
# run tests
make test
# formatting, linting, and type checking
make lint
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
Hashes for siftlog_py-0.9.2-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 594048b208041a77dd6c535c08e0600e293c6182985690e24fb3f4bba3bf19b1 |
|
MD5 | f1a04428fdd660b6ae4a5c1053a8da44 |
|
BLAKE2b-256 | 997936cc000b3cc0260637698959b19f81d5d77d99d4de3b5d4786753b359591 |