봇의 로그 기록과 오류 처리

A bot runs unattended, often on a schedule. When something breaks, you are not watching. Logging leaves a trail so you can diagnose what happened after the fact.

Without logs, a silent failure can go unnoticed for days.

Beginners use print(), but it has no levels, no timestamps, and no easy way to redirect to a file. Python's logging module solves all of this.

Configure logging once at startup. Set a format with a timestamp and a level, and choose where output goes.

Levels let you control verbosity:

• DEBUG detailed tracing.
• INFO normal progress.
• WARNING something odd but recoverable.
• ERROR an operation failed.
• CRITICAL the bot cannot continue.

For unattended bots, write logs to a file so you can review them later. A rotating handler caps file size.

Wrap risky operations in try/except so one failure does not crash the whole run. Log the error with context.

Use log.exception inside an except block to capture the full stack trace, which is invaluable for debugging.

Network blips are common. Catch them and retry a few times before giving up, logging each attempt.

Silent failures are dangerous. For critical errors, escalate: send an alert email, write a marker file, or exit with a non-zero code so a scheduler notices.

Always release resources even on error. A finally block (or context managers) guarantees browsers close and files flush regardless of what went wrong.

Attach context to each log line, such as the current URL or item id, so when you scan a long log you immediately know which record a message refers to.

Test your understanding of bot logging and error handling.

You learned to make simple bots resilient: use the logging module instead of print, choose appropriate levels, log to rotating files, catch and log exceptions with tracebacks, retry transient errors, fail loudly on critical issues, and shut down cleanly.