Skip to content

fix(logger): handle case-insensitive and integer log levels - #8836

Merged
leandrodamascena merged 3 commits into
aws-powertools:developfrom
roonakyadav:fix/8823-log-level-normalization
Nov 20, 2026
Merged

leandrodamascena merged 3 commits into
aws-powertools:developfrom
roonakyadav:fix/8823-log-level-normalization

Conversation

@roonakyadav

Copy link
Copy Markdown
Contributor

Closes #8823

Summary

Changes

  • In aws_lambda_powertools/logging/logger.py:
    • Add _to_numeric_log_level() helper method to resolve string log levels (standard or custom names registered via logging.addLevelName) and integer levels to numeric integers across Python 3.10–3.14 (using logging.getLevelNamesMapping() on Python 3.11+ and falling back to logging._nameToLevel on Python 3.10).
    • Update _get_powertools_log_level() to normalize string log levels case-insensitively (including POWERTOOLS_LOG_LEVEL and LOG_LEVEL environment variables) to uppercase, and preserve valid integer log levels as integers instead of converting them via logging.getLevelName().
    • Update _determine_log_level() to compare numeric levels via _to_numeric_log_level(), avoiding TypeError comparisons between int and str, and return the normalized log level (int or uppercase string) without string conversion.
  • In tests/functional/logger/required_dependencies/test_logger.py:
    • Add parameterized regression test matrix covering lowercase string levels ("debug", "info", "warning", "error", "critical"), custom integer levels (15, 25), and standard environment variables (POWERTOOLS_LOG_LEVEL, LOG_LEVEL), both with and without Advanced Logging Controls (AWS_LAMBDA_LOG_LEVEL).

Root cause

  1. When AWS_LAMBDA_LOG_LEVEL was set, passing a lowercase level like level="debug" caused logging.getLevelName("debug") to return the string "Level debug". Line 1056 then executed logging.getLevelName("INFO") > "Level debug", which attempted to compare int with str, raising TypeError: '>' not supported between instances of 'int' and 'str'.
  2. When a custom integer level like level=15 was passed, _get_powertools_log_level() called logging.getLevelName(15), returning "Level 15". _determine_log_level() then called .upper(), yielding "LEVEL 15". Calling stdlib setLevel("LEVEL 15") failed with ValueError: Unknown level: 'LEVEL 15'.

User experience

  • Before:
    • Logger(level="debug") crashed with TypeError whenever AWS_LAMBDA_LOG_LEVEL was set.
    • Logger(level=15) crashed with ValueError: Unknown level: 'LEVEL 15'.
  • After:
    • Logger(level="debug") works seamlessly with or without AWS_LAMBDA_LOG_LEVEL, respecting ALC precedence and case-insensitivity.
    • Logger(level=15) works seamlessly, preserving integer levels as accepted by stdlib logging.

Verification

I used an AI assistant to prepare this fix and ran all validations locally:

  1. Reproduction & regression verification:
    • Confirmed failure on develop for Logger(level="debug") with ALC (TypeError: '>' not supported between instances of 'int' and 'str') and Logger(level=15) (ValueError: Unknown level: 'LEVEL 15').
    • Ran newly added regression test matrix:
      uv run pytest tests/functional/logger/required_dependencies/test_logger.py -k "test_logger_level_case_insensitive_and_integer or test_logger_env_var_case_insensitive" -v
      -> 23 passed, 89 deselected in 0.14s
  2. Full logger test suite:
    • uv run pytest tests/functional/logger/ tests/unit/logger/
      -> 227 passed, 7 warnings in 0.78s
  3. Linting and formatting:
    • make lint (ruff format + ruff check) -> 1417 files left unchanged, All checks passed!
  4. Static type checking:
    • make mypy -> Success: no issues found in 963 source files
  5. Diff hygiene:
    • git diff --check -> Clean, 0 errors.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

Lowercase log levels crashed with TypeError when AWS_LAMBDA_LOG_LEVEL was
configured due to comparing int with str returned by logging.getLevelName().
Custom integer log levels crashed with ValueError when logging.getLevelName(int)
returned "Level <int>" which failed stdlib setLevel().

Normalize string level names case-insensitively, preserve integer levels,
and compare numeric logging levels. Add regression test matrix.

Closes aws-powertools#8823
@roonakyadav
roonakyadav requested a review from a team as a code owner November 15, 2026 16:10
@powertools-for-aws-oss-automation powertools-for-aws-oss-automation Bot added the size/L Denotes a PR that changes 100-499 lines, ignoring generated files. label Nov 15, 2026
@leandrodamascena leandrodamascena added this to the 3.42.0 milestone Nov 15, 2026

@Manishram-ai Manishram-ai left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for working on this. I tried a few level inputs on develop and on this branch. Lowercase names with AWS_LAMBDA_LOG_LEVEL set and level=15 both work here now. The one difference I found is Logger(level=""), which falls back to POWERTOOLS_LOG_LEVEL or INFO on develop and raises ValueError: Unknown level: '' here. With the two suggestions below, tests/functional/logger and tests/unit/logger pass (228 tests), and ruff and mypy are clean.

Comment thread aws_lambda_powertools/logging/logger.py Outdated
# If level is an int (logging.INFO), return its respective string ("INFO")
if isinstance(level, int):
return logging.getLevelName(level)
if isinstance(level, str):

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The old level or log_level_env sent an empty string on to the env var or INFO. Code like Logger(level=os.getenv("LOG_LEVEL", "")) would start failing on cold start, and checking and level here keeps the old fallback.

Suggested change
if isinstance(level, str):
if isinstance(level, str) and level:

("error", logging.ERROR),
("critical", logging.CRITICAL),
(15, 15),
(logging.INFO, logging.INFO),

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A row for the empty string. It fails on this branch without the change above and passes with it.

Suggested change
(logging.INFO, logging.INFO),
(logging.INFO, logging.INFO),
("", logging.INFO),

@leandrodamascena leandrodamascena left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @roonakyadav! Comparing numbers on both sides instead of mixing names and ints is the right fix, and I like that ints stay ints now. One small note for later, not for this PR: logging.getLevelName(name) already returns an int for known names on every version we support, so the mapping lookup and the private _nameToLevel fallback could go away once 3.10 drops. Goes out in 3.42.0, no need to rebase, you'll be in the release notes.

@sonarqubecloud

Copy link
Copy Markdown

@codecov

codecov Bot commented Nov 20, 2026 •

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 65.38462% with 9 lines in your changes missing coverage. Please review.
✅ Project coverage is 97.15%. Comparing base (ba73859) to head (f5ab2bd).

Files with missing lines Patch % Lines
aws_lambda_powertools/logging/logger.py 65.38% 7 Missing and 2 partials ⚠️
Additional details and impacted files
@@             Coverage Diff             @@
##           develop    #8836      +/-   ##
===========================================
- Coverage    97.20%   97.15%   -0.06%     
===========================================
  Files          327      327              
  Lines        16761    16780      +19     
  Branches      1523     1530       +7     
===========================================
+ Hits         16292    16302      +10     
- Misses         328      335       +7     
- Partials       141      143       +2     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@leandrodamascena
leandrodamascena merged commit 8a21c19 into aws-powertools:develop Nov 20, 2026
17 of 20 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

logger size/L Denotes a PR that changes 100-499 lines, ignoring generated files. tests

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Bug: lowercase log level with Advanced Logging Controls crashes, and custom int levels crash

3 participants