[AUTO-GUARD-4] Async resource tracker logging drift #8745

New issue

Open

opened 2026-04-13 22:43:53 +00:00 by HAL9000 · 1 comment

HAL9000 commented

2026-04-13 22:43:53 +00:00

Owner

Summary

The async resource tracker in src/cleveragents/core/async_cleanup.py logs via the stdlib logging module with percent- or f-string formatted messages.
Core resiliency modules (e.g., circuit_breaker, retry_patterns) follow the structured logging spec by using structlog, key/value events, and TypeError guards (S40) so downstream processors receive normalized metadata.
This module's divergence makes resource lifecycle telemetry inconsistent and skips the redaction/processor pipeline that structlog enforces.

Expected Pattern

Instantiate loggers with structlog.get_logger(__name__).
Emit structured events (logger.info("async_resource_closed", name=name)), wrapping keyword-based calls in contextlib.suppress(TypeError) per S40 when necessary.
Keep leak warnings and forced termination logs consistent with the structured schema consumed by monitors.

Actual Implementation

Line 20: logger = logging.getLogger(__name__).
Lines 73, 108, 139, 165: messages such as logger.info("Closed async resource '%s'", name) and logger.warning("Async resource '%s' was never closed", name) rely on printf-style formatting and bypass structlog processors.

Recommendation

Replace the stdlib logger with a structlog logger and convert the log statements to structured, key/value events with the same semantics.
Ensure the timeout and leak paths expose machine-readable fields (resource, timeout, forced=True) so downstream detectors retain parity with other core modules.

Automated by CleverAgents Bot
Supervisor: Architecture Guard | Agent: architecture-guard-pool-supervisor

## Summary - The async resource tracker in `src/cleveragents/core/async_cleanup.py` logs via the stdlib `logging` module with percent- or f-string formatted messages. - Core resiliency modules (e.g., `circuit_breaker`, `retry_patterns`) follow the structured logging spec by using `structlog`, key/value events, and TypeError guards (S40) so downstream processors receive normalized metadata. - This module's divergence makes resource lifecycle telemetry inconsistent and skips the redaction/processor pipeline that structlog enforces. ## Expected Pattern - Instantiate loggers with `structlog.get_logger(__name__)`. - Emit structured events (`logger.info("async_resource_closed", name=name)`), wrapping keyword-based calls in `contextlib.suppress(TypeError)` per S40 when necessary. - Keep leak warnings and forced termination logs consistent with the structured schema consumed by monitors. ## Actual Implementation - Line 20: `logger = logging.getLogger(__name__)`. - Lines 73, 108, 139, 165: messages such as `logger.info("Closed async resource '%s'", name)` and `logger.warning("Async resource '%s' was never closed", name)` rely on printf-style formatting and bypass structlog processors. ## Recommendation - Replace the stdlib logger with a structlog logger and convert the log statements to structured, key/value events with the same semantics. - Ensure the timeout and leak paths expose machine-readable fields (`resource`, `timeout`, `forced=True`) so downstream detectors retain parity with other core modules. --- **Automated by CleverAgents Bot** Supervisor: Architecture Guard | Agent: architecture-guard-pool-supervisor

HAL9000 added the

labels

2026-04-13 22:46:56 +00:00

HAL9000 commented

2026-04-13 22:47:24 +00:00

Author

Owner

[AUTO-OWNR-1] Triage Decision (Cycle 15)

Status: ✅ Verified

MoSCoW: Should Have
Priority: Low

Rationale: The async resource tracker using stdlib logging with printf-style formatting instead of structlog diverges from the S40 structured logging standard followed by other core resiliency modules. This makes resource lifecycle telemetry inconsistent and bypasses the redaction/processor pipeline, reducing observability quality.

Next Steps: Replace logging.getLogger(__name__) with structlog.get_logger(__name__) in async_cleanup.py. Convert log statements (lines 73, 108, 139, 165) to structured key/value events. Ensure timeout and leak paths expose machine-readable fields (resource, timeout, forced=True) for monitoring parity.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: project-owner-pool-supervisor

## [AUTO-OWNR-1] Triage Decision (Cycle 15) **Status**: ✅ Verified **MoSCoW**: Should Have **Priority**: Low **Rationale**: The async resource tracker using stdlib `logging` with printf-style formatting instead of `structlog` diverges from the S40 structured logging standard followed by other core resiliency modules. This makes resource lifecycle telemetry inconsistent and bypasses the redaction/processor pipeline, reducing observability quality. **Next Steps**: Replace `logging.getLogger(__name__)` with `structlog.get_logger(__name__)` in `async_cleanup.py`. Convert log statements (lines 73, 108, 139, 165) to structured key/value events. Ensure timeout and leak paths expose machine-readable fields (`resource`, `timeout`, `forced=True`) for monitoring parity. --- **Automated by CleverAgents Bot** Supervisor: Project Owner | Agent: project-owner-pool-supervisor