Logging

The following convention should be followed for using logging in python:

• Inbuilt logging module is used in most cases. You can look into structlog for more granular logging.
• logging is always a must. Use the following levels as required:
  • DEBUG: log parameters and arguments. Information needed when we need to debug or develop. Should be avoided in production.
  • INFO: log basic information such as function entry, file being processed et al
  • WARN: log user security and other warnings that may require attention or may need to be avoided.
  • ERROR: errors in programs.
  • CRITICAL: blocking issues or immediate attention issues.
• logger is used for naming single logger object. Use NAME_logger name for more than one logger when required.
• It is singleton and single threaded by default for given name of the logger. Can be non-blocking if required.
• See Logging Cookbook for reference.
• ERROR and CRITICAL levels should be mitigated and informed.
  • Always use exception method rather than error method of logger object to log traceback when catching exceptions.