client.logging

Logging configuration for jobmon client applications.

This module provides the standard logging configuration used by workflow.run(configure_logging=True).

The module supports both legacy dict-based configs and new template-based configurations. Requester logs are automatically captured by OTLP when enabled (handled by the Requester class itself).

Attributes

_structlog_configured_lock
_structlog_configured_by_jobmon
_DEFAULT_LOG_FORMAT
default_config

Functions

ensure_structlog_configured(→ None)	Ensure structlog is configured for Jobmon, using lazy initialization.
configure_client_logging(→ None)	Configure client logging with template and user override support.
_remove_non_jobmon_handlers(→ List[str])	Keep only Jobmon OTLP handlers when host renders output directly.
_configure_client_logging_for_direct_rendering(→ None)	Configure logging for direct-rendering hosts while preserving telemetry.
_ensure_jobmon_otlp_handlers(→ List[str])	Attach OTLP handlers if jobmon loggers ended up without handlers.

Module Contents

client.logging._structlog_configured_lock

client.logging._structlog_configured_by_jobmon = False

client.logging._DEFAULT_LOG_FORMAT = '%(asctime)s [%(name)-12s] %(module)s %(levelname)-8s: %(message)s'

client.logging.default_config: Dict

client.logging.ensure_structlog_configured() → None

Ensure structlog is configured for Jobmon, using lazy initialization.

This function is called at the start of key Jobmon operations (workflow.run(), workflow.bind(), etc.) to ensure structlog is configured. It uses a lock to ensure thread-safe initialization.

Behavior: - If host has already configured structlog: prepend Jobmon processors - If structlog is unconfigured: configure it with Jobmon defaults - If already called: no-op (idempotent)

This lazy approach ensures host applications always have the opportunity to configure structlog first, eliminating import-order dependencies.

client.logging.configure_client_logging() → None

Configure client logging with template and user override support.

This is the primary interface for configuring client logging. It supports: 1. Default template-based configuration 2. User file overrides via logging.client_logconfig_file 3. User section overrides via logging.client.* 4. Environment variable overrides

Configuration precedence: 1. Custom file (logging.client_logconfig_file) 2. Section overrides (logging.client.formatters/handlers/loggers) 3. Default template (logconfig_client.yaml) 4. Basic fallback configuration

Adapts to host application logging architecture: - If host uses direct rendering (like FHS): sets up minimal stdlib handlers

that pass through formatted output without double-processing

• If host uses stdlib integration: sets up full Jobmon logging configuration

Note: Requester OTLP is handled separately by the Requester class.

client.logging._remove_non_jobmon_handlers(logconfig: Dict) → List[str]

Keep only Jobmon OTLP handlers when host renders output directly.

client.logging._configure_client_logging_for_direct_rendering(*, default_template_path: str, logger_factory: Any, wrapper_class: Any) → None

Configure logging for direct-rendering hosts while preserving telemetry.

client.logging._ensure_jobmon_otlp_handlers(logconfig_data: Dict[str, Any], jobmon_logger_names: Iterable[str] | None = None) → List[str]

Attach OTLP handlers if jobmon loggers ended up without handlers.