Skip to content

Policies

loly enforces logging best practices through policies. Think of them as guardrails between your code and the 3 AM incident that will haunt your dreams.

Active Policies

These aren't suggestions. They're rules. Break them at your own peril.

Code Name Severity Your Fate
LY001 Exception exc_info Critical No stack traces = No sleep
LY002 Log Loop Critical Spammy logs = Angry ops teams

Coming Soon (To Save You)

These are in development because apparently, nobody listens until they're already broken.

Code Name Impact ETA
LY003 Lazy Logging High - Your performance depends on it Soon™
LY004 Safe Serialization Medium - Prevents explosion in production Soon™
LY005 Logger Naming Medium - Better filtering, better debugging Soon™

Policy Philosophy

We believe in a simple approach:

Actionable: Every violation has a crystal-clear fix. No ambiguity. No excuses.

Zero false positives: We only flag genuine, production-grade problems. Not perfect, but real.

Configurable: Strict in CI, gentle in local development. Your choice.

Performance-focused: Static analysis only—we're not here to slow you down. Just stop you from breaking things.

How Policies Work

Each policy uses Python's AST (Abstract Syntax Tree) to analyze code. No regex nonsense here.

  1. Parse Python files with libcst - the right way
  2. Walk the AST looking for patterns - precise and unforgiving
  3. Flag violations with file, line, and message - so you know exactly what broke
  4. Suggest fixes - because we're not here to ruin your day, just your mistakes

This approach is:

  • Fast: Analyzes thousands of files in seconds. Even your disaster codebase.
  • Accurate: AST-based analysis means zero false positives. We're not guessing.
  • Safe: Read-only. We find problems, we don't break your code.

Policy Configuration

Customize each policy in loly.yml to match your risk tolerance:

exception_exc_info:
  levels: [error]        # Only check error logs (recommended)
  severity: fail         # Block CI. No mercy.

log_loop:
  levels: [info, debug]  # Check info and debug logs
  severity: warn         # Warn but don't block (easing in? Fair enough.)

Severity Levels

fail: Exit code 1, blocks CI. This is the serious setting.

warn: Print a warning, exit code 0. The "I'll fix it later" setting. (You won't.)

info: Just informational. Perfect if you like living dangerously.

Ready to Learn?