Python - 如何在所有日志级别甚至外部异常打印堆栈跟踪？

编程问答 2022-04-28

问题描述

我的代码库非常大，我们想打印所有日志的堆栈。

这包括所有级别的日志（甚至 INFO 和 DEBUG）。

这还包括发生异常之外的日志。

我学到的东西：

  • 除非你在一个except中，否则不会显示堆栈跟踪：调用logger.exception时，set_exec=True，trace back.print_exc()
  • 我发现显示堆栈的唯一方法是 traceback.print_stack()
  • 我现在希望我的所有日​​志都打印此内容，但我不知道如何覆盖 logger.debug/info/warning/error 以添加 traceback.print_stack()
  • 这可以使用格式化程序吗？如果没有，是否可以在发送到格式化程序之前覆盖 logger.info() 方法？
  • 还有其他建议吗？

解决方法

您走在正确的轨道上。我建议您阅读logging HOWTO，它解释了很多事情。

您可以使用级别和过滤器来决定您的处理程序是否应该处理日志。您可以使用格式化程序将堆栈跟踪添加到日志记录消息中。

这是一个概念验证：

import logging
import traceback
import sys


logger = logging.getLogger("app.main")


class StacktraceLogFormatter(logging.Formatter):
    def format(self,record: logging.LogRecord) -> str:
        stack_lines = traceback.format_stack()
        # we have the lines down to ^^^^^^ this call,but we don't want to show it,nor the internal logging ones
        stack_lines_without_logging_intermediary_calls = filter(
            lambda line: ("lib/logging/__init__.py" not in line)
                         and ("lib\\logging\\__init__.py") not in line,stack_lines[:-1]
        )
        return record.msg + "\nOrigin :\n" + "".join(stack_lines_without_logging_intermediary_calls)
        # beware of : https://stackoverflow.com/q/8162419/11384184


class InfoOrLessFilter(logging.Filter):
    def filter(self,record: logging.LogRecord) -> bool:
        return record.levelno <= logging.INFO


def do_some_logging():
    logger.debug("debug message")
    logger.info("info message")
    logger.warning("warning message")
    logger.error("error message")
    try:
        1/0
    except ZeroDivisionError:
        logger.exception("exception message")
    logger.critical("critical message")


def setup_custom_logging():
    logger.handlers = []  # remove the existing handlers from the logger

    regular_handler_for_info_messages = logging.StreamHandler(sys.stdout)
    regular_handler_for_info_messages.setLevel(logging.DEBUG)  # at least DEBUG
    regular_handler_for_info_messages.addFilter(InfoOrLessFilter())  # but not more than INFO
    logger.addHandler(regular_handler_for_info_messages)

    stacktrace_handler_for_important_messages = logging.StreamHandler(sys.stderr)
    stacktrace_handler_for_important_messages.setLevel(logging.INFO + 1)  # more than INFO
    stacktrace_handler_for_important_messages.setFormatter(StacktraceLogFormatter())
    logger.addHandler(stacktrace_handler_for_important_messages)

    logger.propagate = False


def main():
    logging.basicConfig(level=logging.DEBUG)
    do_some_logging()
    setup_custom_logging()
    do_some_logging()


if __name__ == "__main__":
    main()

它之前产生：

debug message
info message
DEBUG:app.main:debug message
INFO:app.main:info message
WARNING:app.main:warning message
ERROR:app.main:error message
ERROR:app.main:exception message
Traceback (most recent call last):
  File "C:/PycharmProjects/stack_overflow/67846424.py",line 43,in do_some_logging
    1/0
ZeroDivisionError: division by zero
CRITICAL:app.main:critical message

之后：

warning message
Origin :
  File "C:/PycharmProjects/stack_overflow/67846424.py",line 73,in <module>
    main()
  File "C:/PycharmProjects/stack_overflow/67846424.py",line 69,in main
    do_some_logging()
  File "C:/PycharmProjects/stack_overflow/67846424.py",line 40,in do_some_logging
    logger.warning("warning message")

error message
Origin :
  File "C:/PycharmProjects/stack_overflow/67846424.py",line 41,in do_some_logging
    logger.error("error message")

exception message
Origin :
  File "C:/PycharmProjects/stack_overflow/67846424.py",line 45,in do_some_logging
    logger.exception("exception message")

critical message
Origin :
  File "C:/PycharmProjects/stack_overflow/67846424.py",line 46,in do_some_logging
    logger.critical("critical message")
formatterloggingpythonstack-trace