Supprimer Rendre public Rendre privé Add tags Delete tags
  Ajouter un tag   Annuler
  Supprimer le tag   Annuler
  • • DevOps notes •
  •  
  • Tags
  • Connexion

Log Levels in Practice/shaare/sMZ4dg

  • python
  • python

Log Levels in Practice

  • Python defines five standard levels with increasing severity:
    • DEBUG (10): Detailed diagnostic information.
    • INFO (20): Confirmation that things are working normally.
    • WARNING (30): An indication of potential problems or deprecation.
    • ERROR (40): A failure in a specific operation.
    • CRITICAL (50): A serious error causing program termination.
  • NOTSET (0) causes a logger to inherit its parent’s effective level.
  • Appropriate use of these levels lets you adjust verbosity without changing code.

Two-Stage Filtering: Logger vs Handler

  • Logger Level: First gate: records below logger.level are discarded immediately.
  • Handler Level: Second gate: each handler only emits records at or above its handler.level.
  • This allows, for example, DEBUG messages to be logged to a file but only WARNING and above to the console.

Configuring Logger & Handlers

  • Use logger.setLevel(...) to control which messages the logger accepts.
  • Use handler.setLevel(...) to control which accepted messages each handler emits.
  • Attach multiple handlers for different ouputs (e.g., console vs file) with independent levels.

# Log levels in practice

import logging
import sys

print("Log levels in practice")
print("------\n")

for lvl in (
    logging.DEBUG,
    logging.INFO,
    logging.WARNING,
    logging.ERROR,
    logging.CRITICAL,
):
    print(
        f"{logging.getLevelName(lvl):8} = {lvl}"
    )

# Two-stage filtering

print("\n")
print("Two stage filtering")
print("------\n")

filter_logger = logging.getLogger("demo.filter")
filter_logger.setLevel(logging.INFO)

stream_handler = logging.StreamHandler(sys.stdout)
stream_handler.setLevel(logging.ERROR)

filter_logger.addHandler(stream_handler)

filter_logger.info("INFO: will not be shown")
filter_logger.error("ERROR: will be shown")

# Configuring logs and handlers

print("\n")
print("Configuring logs and handlers")
print("------\n")

data_logger = logging.getLogger("demo.data")
data_logger.setLevel(logging.DEBUG)

data_sh = logging.StreamHandler(sys.stdout)
data_sh.setLevel(logging.ERROR)

data_fh = logging.FileHandler("process.log", "w")
data_fh.setLevel(logging.INFO)

data_logger.addHandler(data_sh)
data_logger.addHandler(data_fh)

data_logger.debug("DEBUG: will be dropped")
data_logger.info("INFO: file only")
data_logger.warning("WARNING: file only")
data_logger.error("ERROR: file and console")
data_logger.critical("CRITICAL: file and console")
4 months ago Permalien
cluster icon
  • Filesystem Operations : Filesystem Operations (os & shutil) DevOps scripts often need to create, delete, copy, and move files and directories as part of automation workflows...
  • Generators and Lazy Pipelines : Generators and Lazy Pipelines You can chain generator functions to form multi-stage data pipelines that process items one at a time. No intermediat...
  • Adding Type Hints to Decorators and Generators : Adding Type Hints to Decorators and Generators Decorators and generators are advanced constructs that require specialized type hints to make their tr...
  • Structured Logging : Introduction to Structured Logging Plain-text logs are hard to parse and brittle to format changes. Structured logging records events as key-value da...
  • Python package and subpackage : Introduction to Packages (__init__.py) What is a Package? A Python package provides a way to structure a project's module namespace by using directori...


(110)
Filtrer par liens sans tag
Replier Replier tout Déplier Déplier tout Êtes-vous sûr de vouloir supprimer ce lien ? Êtes-vous sûr de vouloir supprimer ce tag ? Le gestionnaire de marque-pages personnel, minimaliste, et sans base de données par la communauté Shaarli