omnibase_infra 0.2.6__py3-none-any.whl

omnibase_infra/observability/handlers/handler_logging_structured.py

@@ -0,0 +1,967 @@
+# SPDX-License-Identifier: MIT
+# Copyright (c) 2025 OmniNode Team
+"""Structured Logging Handler - EFFECT handler for structured logging with buffer/flush.
+
+This handler implements the ONEX EFFECT handler pattern for structured logging.
+It manages the lifecycle of a SinkLoggingStructured instance, providing:
+- Contract-driven lifecycle management (initialize, shutdown)
+- Periodic background flush with configurable interval
+- Buffer threshold flush (when buffer reaches capacity)
+- Graceful shutdown with final flush
+
+Architecture Principle: "Handlers own lifecycle, sinks own hot path"
+- This handler: Contract-driven lifecycle, buffer/flush management
+- SinkLoggingStructured: Fast in-process emission (synchronous, non-blocking)
+
+Supported Operations:
+    - logging.emit: Emit a log entry to the buffer (delegates to sink)
+    - logging.flush: Force flush all buffered entries immediately
+    - logging.configure: Update logging configuration at runtime
+
+Thread Safety:
+    The handler uses a multi-lock design for safe concurrent access:
+
+    1. _config_lock (asyncio.Lock): Guards async configuration changes and prevents
+       concurrent configure operations from interfering with each other.
+
+    2. _emit_lock (threading.Lock): Guards the critical section during emit and
+       reconfiguration to prevent log loss. This lock ensures that:
+       - While an emit is in progress, reconfiguration waits for it to complete
+       - While reconfiguration is in progress (swap + flush), emits are serialized
+
+    The underlying sink uses its own threading.Lock for buffer operations.
+    This multi-lock design allows safe operation from both async and sync contexts
+    while preventing log loss during reconfiguration.
+
+Reconfiguration Log Loss Prevention:
+    Without synchronization, a race condition can cause log loss:
+    1. emit reads self._sink -> gets old_sink reference
+    2. configure swaps self._sink to new_sink
+    3. emit calls old_sink.emit() -> log goes to abandoned sink (LOST)
+
+    The _emit_lock prevents this by ensuring the swap is atomic with respect
+    to emit operations. The swap-then-flush sequence is:
+    1. Acquire _emit_lock (blocks all emits)
+    2. Swap to new sink (atomic reference swap)
+    3. Release _emit_lock (emits resume to new sink immediately)
+    4. Flush old sink (outside lock - logs during flush go to new sink)
+
+    This guarantees no log loss while minimizing lock hold time. The old sink's
+    buffer is preserved until explicitly flushed, and any logs emitted during
+    the flush operation go directly to the new sink.
+
+Drop Policy:
+    The drop_policy configuration field accepts only "drop_oldest", which is the
+    sole supported policy. When the buffer is full, the oldest entries are dropped
+    to make room for new ones, preserving recent logs. The policy is applied to
+    the sink during both initialization and reconfiguration.
+
+Envelope-Based Routing:
+    This handler uses envelope-based operation routing. See CLAUDE.md section
+    "Intent Model Architecture > Envelope-Based Handler Routing" for the full
+    design pattern.
+"""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+import threading
+from typing import TYPE_CHECKING
+from uuid import UUID, uuid4
+
+from omnibase_core.types import JsonType
+from omnibase_infra.enums import (
+    EnumHandlerType,
+    EnumHandlerTypeCategory,
+    EnumInfraTransportType,
+    EnumResponseStatus,
+)
+from omnibase_infra.errors import (
+    ModelInfraErrorContext,
+    ProtocolConfigurationError,
+    RuntimeHostError,
+)
+from omnibase_infra.mixins import MixinEnvelopeExtraction
+from omnibase_infra.observability.handlers.model_logging_handler_config import (
+    ModelLoggingHandlerConfig,
+)
+from omnibase_infra.observability.handlers.model_logging_handler_response import (
+    ModelLoggingHandlerResponse,
+)
+
+# Dependency validation for structlog (via SinkLoggingStructured)
+# The sink requires structlog; check availability here for early error detection
+_STRUCTLOG_AVAILABLE: bool = False
+try:
+    import structlog
+
+    _STRUCTLOG_AVAILABLE = True
+    del structlog  # Not used at runtime, only for dependency check
+except ImportError:
+    pass
+
+# Import sink after dependency check (will fail at instantiation if structlog missing)
+from omnibase_infra.observability.sinks import SinkLoggingStructured
+
+if TYPE_CHECKING:
+    from omnibase_core.enums import EnumLogLevel
+
+logger = logging.getLogger(__name__)
+
+# Handler ID for responses
+HANDLER_ID_LOGGING: str = "logging-handler"
+
+# Supported operations
+SUPPORTED_OPERATIONS: frozenset[str] = frozenset(
+    {
+        "logging.emit",
+        "logging.flush",
+        "logging.configure",
+    }
+)
+
+
+class HandlerLoggingStructured(MixinEnvelopeExtraction):
+    """Structured logging EFFECT handler with buffer/flush management.
+
+    This handler manages the lifecycle of a SinkLoggingStructured instance,
+    providing contract-driven initialization, periodic background flush,
+    and graceful shutdown with final flush.
+
+    Lifecycle:
+        1. initialize(): Creates sink, starts periodic flush task
+        2. execute(): Routes operations to sink methods
+        3. shutdown(): Cancels flush task, performs final flush
+
+    Buffer Management:
+        - The sink maintains a bounded buffer with configurable size
+        - Periodic flush: Background task flushes at configured interval
+        - Threshold flush: Automatic flush when buffer nears capacity
+        - Shutdown flush: Final flush ensures no data loss
+
+    Buffer Metrics:
+        The periodic flush loop emits the following metrics for monitoring:
+        - logging_buffer_utilization_ratio: Gauge (0.0-1.0) of buffer usage
+        - logging_buffer_dropped_total: Counter of entries dropped since last flush
+
+        These metrics help detect if buffer size is insufficient before logs are lost.
+
+    Thread Safety:
+        - Configuration changes use asyncio.Lock (prevents concurrent configure calls)
+        - Emit operations use threading.Lock (prevents log loss during reconfiguration)
+        - Sink operations use threading.Lock (internal to sink for buffer access)
+        - Safe for concurrent async callers; emit is protected during sink swap
+
+    Example:
+        ```python
+        handler = HandlerLoggingStructured()
+        await handler.initialize({
+            "buffer_size": 500,
+            "flush_interval_seconds": 10.0,
+            "output_format": "json",
+        })
+
+        # Emit log entries
+        await handler.execute({
+            "operation": "logging.emit",
+            "payload": {
+                "level": "INFO",
+                "message": "User logged in",
+                "context": {"user_id": "u_123"},
+            },
+        })
+
+        # Force flush
+        await handler.execute({
+            "operation": "logging.flush",
+            "payload": {},
+        })
+
+        await handler.shutdown()
+        ```
+    """
+
+    def __init__(self) -> None:
+        """Initialize handler in uninitialized state."""
+        self._sink: SinkLoggingStructured | None = None
+        self._config: ModelLoggingHandlerConfig | None = None
+        self._initialized: bool = False
+        self._flush_task: asyncio.Task[None] | None = None
+        self._shutdown_event: asyncio.Event | None = None
+        self._config_lock: asyncio.Lock = asyncio.Lock()
+        # Threading lock to prevent log loss during reconfiguration.
+        # This lock ensures emit operations complete before sink swap+flush.
+        self._emit_lock: threading.Lock = threading.Lock()
+        # Track dropped count from previous flush cycle to emit delta
+        self._last_drop_count: int = 0
+
+    @property
+    def handler_type(self) -> EnumHandlerType:
+        """Return the architectural role of this handler.
+
+        Returns:
+            EnumHandlerType.INFRA_HANDLER - This handler is an infrastructure
+            handler that manages the logging sink lifecycle. It is not a
+            NODE_HANDLER (event processing) or COMPUTE_HANDLER (pure computation).
+
+        Note:
+            handler_type determines lifecycle, protocol selection, and runtime
+            invocation patterns. It answers "what is this handler in the architecture?"
+
+        See Also:
+            - handler_category: Behavioral classification (EFFECT/COMPUTE)
+            - docs/architecture/HANDLER_PROTOCOL_DRIVEN_ARCHITECTURE.md
+        """
+        return EnumHandlerType.INFRA_HANDLER
+
+    @property
+    def handler_category(self) -> EnumHandlerTypeCategory:
+        """Return the behavioral classification of this handler.
+
+        Returns:
+            EnumHandlerTypeCategory.EFFECT - This handler performs side-effecting
+            I/O operations (writing logs to stdout/files). EFFECT handlers are
+            not deterministic and interact with external systems.
+
+        Note:
+            handler_category determines security rules, determinism guarantees,
+            replay safety, and permissions. It answers "how does this handler
+            behave at runtime?"
+
+            Categories:
+            - COMPUTE: Pure, deterministic transformations (no side effects)
+            - EFFECT: Side-effecting I/O (database, HTTP, file writes, logging)
+            - NONDETERMINISTIC_COMPUTE: Pure but not deterministic (UUID, random)
+
+        See Also:
+            - handler_type: Architectural role (INFRA_HANDLER/NODE_HANDLER/etc.)
+            - docs/architecture/HANDLER_PROTOCOL_DRIVEN_ARCHITECTURE.md
+        """
+        return EnumHandlerTypeCategory.EFFECT
+
+    async def initialize(self, config: dict[str, object]) -> None:
+        """Initialize the logging handler with configuration.
+
+        Creates the SinkLoggingStructured instance and starts the periodic
+        flush background task if configured.
+
+        Args:
+            config: Configuration dict containing:
+                - buffer_size: Max buffer size (default: 1000)
+                - flush_interval_seconds: Flush interval (default: 5.0, 0 to disable)
+                - output_format: "json" or "console" (default: "json")
+                - drop_policy: "drop_oldest" (only supported policy)
+
+        Raises:
+            ProtocolConfigurationError: If structlog is not installed or
+                configuration validation fails.
+            RuntimeHostError: If handler is already initialized.
+        """
+        init_correlation_id = uuid4()
+
+        # Validate required dependencies before proceeding
+        if not _STRUCTLOG_AVAILABLE:
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=init_correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation="initialize",
+                target_name="logging_handler",
+            )
+            raise ProtocolConfigurationError(
+                "Missing required dependency: structlog. "
+                "Install with: pip install structlog",
+                context=ctx,
+            )
+
+        logger.info(
+            "Initializing %s",
+            self.__class__.__name__,
+            extra={
+                "handler": self.__class__.__name__,
+                "correlation_id": str(init_correlation_id),
+            },
+        )
+
+        if self._initialized:
+            # Log warning before raising to help users understand their config
+            # is not being applied (the handler maintains existing configuration)
+            logger.warning(
+                "Handler %s already initialized; ignoring new configuration. "
+                "Call shutdown() first to reinitialize with new config.",
+                self.__class__.__name__,
+                extra={
+                    "handler": self.__class__.__name__,
+                    "correlation_id": str(init_correlation_id),
+                    "existing_config": self._config.model_dump()
+                    if self._config
+                    else None,
+                },
+            )
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=init_correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation="initialize",
+                target_name="logging_handler",
+            )
+            raise RuntimeHostError(
+                "Handler already initialized. Call shutdown() first.",
+                context=ctx,
+            )
+
+        # Validate and parse configuration
+        # NOTE: Broad Exception catch is intentional here because Pydantic can raise
+        # various exception types (ValidationError, TypeError, ValueError) depending
+        # on the validation failure. We wrap all in ProtocolConfigurationError.
+        try:
+            self._config = ModelLoggingHandlerConfig.model_validate(config)
+        except Exception as e:
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=init_correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation="initialize",
+                target_name="logging_handler",
+            )
+            raise ProtocolConfigurationError(
+                f"Invalid logging handler configuration: {e}",
+                context=ctx,
+            ) from e
+
+        # Create the sink with ALL config fields applied, including drop_policy.
+        # The drop_policy controls buffer overflow behavior (currently only "drop_oldest").
+        self._sink = SinkLoggingStructured(
+            max_buffer_size=self._config.buffer_size,
+            output_format=self._config.output_format,
+            drop_policy=self._config.drop_policy,
+        )
+
+        # Reset drop count tracker for new sink
+        self._last_drop_count = 0
+
+        # Set up shutdown event and start periodic flush task
+        self._shutdown_event = asyncio.Event()
+
+        if self._config.flush_interval_seconds > 0:
+            self._flush_task = asyncio.create_task(
+                self._periodic_flush_loop(),
+                name="logging-handler-flush",
+            )
+
+        self._initialized = True
+
+        logger.info(
+            "%s initialized successfully",
+            self.__class__.__name__,
+            extra={
+                "handler": self.__class__.__name__,
+                "buffer_size": self._config.buffer_size,
+                "flush_interval_seconds": self._config.flush_interval_seconds,
+                "output_format": self._config.output_format,
+                "drop_policy": self._config.drop_policy,
+                "correlation_id": str(init_correlation_id),
+            },
+        )
+
+    async def shutdown(self) -> None:
+        """Shutdown the logging handler gracefully.
+
+        Cancels the periodic flush task, performs a final flush of all
+        buffered entries, and releases resources.
+
+        This method is idempotent - calling it multiple times is safe.
+        """
+        shutdown_correlation_id = uuid4()
+
+        logger.info(
+            "Shutting down %s",
+            self.__class__.__name__,
+            extra={
+                "handler": self.__class__.__name__,
+                "correlation_id": str(shutdown_correlation_id),
+            },
+        )
+
+        # Signal shutdown to flush loop
+        if self._shutdown_event is not None:
+            self._shutdown_event.set()
+
+        # Cancel periodic flush task
+        if self._flush_task is not None:
+            self._flush_task.cancel()
+            try:
+                await self._flush_task
+            except asyncio.CancelledError:
+                pass
+            self._flush_task = None
+
+        # Final flush - best effort, never fails shutdown
+        # NOTE: Broad Exception catch is intentional here to ensure shutdown
+        # always completes regardless of I/O errors. Any flush error during
+        # shutdown is logged but does not prevent cleanup from completing.
+        if self._sink is not None:
+            try:
+                self._sink.flush()
+            except Exception as e:
+                logger.warning(
+                    "Error during final flush on shutdown: %s",
+                    e,
+                    extra={
+                        "handler": self.__class__.__name__,
+                        "correlation_id": str(shutdown_correlation_id),
+                        "error_type": type(e).__name__,
+                    },
+                )
+
+        # Clear state
+        self._sink = None
+        self._config = None
+        self._initialized = False
+        self._shutdown_event = None
+
+        logger.info(
+            "%s shutdown complete",
+            self.__class__.__name__,
+            extra={
+                "handler": self.__class__.__name__,
+                "correlation_id": str(shutdown_correlation_id),
+            },
+        )
+
+    async def execute(self, envelope: dict[str, object]) -> ModelLoggingHandlerResponse:
+        """Execute a logging operation from the envelope.
+
+        Supported operations:
+            - logging.emit: Buffer a log entry
+            - logging.flush: Force flush all buffered entries
+            - logging.configure: Update handler configuration
+
+        Args:
+            envelope: Request envelope containing:
+                - operation: Operation identifier (logging.emit, etc.)
+                - payload: Operation-specific payload
+                - correlation_id: Optional correlation ID for tracing
+                - envelope_id: Optional envelope ID for causality tracking
+
+        Returns:
+            ModelLoggingHandlerResponse with operation result
+
+        Raises:
+            RuntimeHostError: If handler not initialized or invalid operation.
+        """
+        correlation_id = self._extract_correlation_id(envelope)
+
+        if not self._initialized or self._sink is None or self._config is None:
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation="execute",
+                target_name="logging_handler",
+            )
+            raise RuntimeHostError(
+                "Logging handler not initialized. Call initialize() first.",
+                context=ctx,
+            )
+
+        operation = envelope.get("operation")
+        if not isinstance(operation, str):
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation="execute",
+                target_name="logging_handler",
+            )
+            raise RuntimeHostError(
+                "Missing or invalid 'operation' in envelope",
+                context=ctx,
+            )
+
+        if operation not in SUPPORTED_OPERATIONS:
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation=operation,
+                target_name="logging_handler",
+            )
+            raise RuntimeHostError(
+                f"Operation '{operation}' not supported. "
+                f"Available: {', '.join(sorted(SUPPORTED_OPERATIONS))}",
+                context=ctx,
+            )
+
+        payload = envelope.get("payload")
+        if not isinstance(payload, dict):
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation=operation,
+                target_name="logging_handler",
+            )
+            raise RuntimeHostError(
+                "Missing or invalid 'payload' in envelope",
+                context=ctx,
+            )
+
+        # Route to operation handler
+        if operation == "logging.emit":
+            return self._handle_emit(payload, correlation_id)
+        elif operation == "logging.flush":
+            return self._handle_flush(correlation_id)
+        else:  # logging.configure
+            return await self._handle_configure(payload, correlation_id)
+
+    def _handle_emit(
+        self,
+        payload: dict[str, object],
+        correlation_id: UUID,
+    ) -> ModelLoggingHandlerResponse:
+        """Handle logging.emit operation.
+
+        Args:
+            payload: Payload containing:
+                - level: Log level (DEBUG, INFO, WARNING, ERROR, CRITICAL, etc.)
+                - message: Log message string
+                - context: Dict of string key-value context data
+            correlation_id: Correlation ID for tracing
+
+        Returns:
+            ModelLoggingHandlerResponse with operation result
+
+        Thread Safety:
+            This method acquires _emit_lock to prevent log loss during
+            reconfiguration. The lock ensures the sink reference is stable
+            throughout the emit operation.
+        """
+        # Extract and validate payload fields BEFORE acquiring lock
+        # to minimize time spent holding the lock
+        level_raw = payload.get("level")
+        message = payload.get("message")
+        context_raw = payload.get("context", {})
+
+        if not isinstance(message, str):
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation="logging.emit",
+                target_name="logging_handler",
+            )
+            raise RuntimeHostError(
+                "Missing or invalid 'message' in payload - must be string",
+                context=ctx,
+            )
+
+        if not isinstance(context_raw, dict):
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation="logging.emit",
+                target_name="logging_handler",
+            )
+            raise RuntimeHostError(
+                "Invalid 'context' in payload - must be dict",
+                context=ctx,
+            )
+
+        # Pass through JSON-compatible context values (sink accepts JsonType)
+        # This preserves rich context data (int, float, bool, list, dict) for structured logging
+        context: dict[str, JsonType] = {str(k): v for k, v in context_raw.items()}
+
+        # Parse log level
+        level = self._parse_log_level(level_raw, correlation_id)
+
+        # Critical section: acquire emit lock to prevent log loss during reconfiguration.
+        # This ensures the sink reference is stable and any logs emitted here will be
+        # flushed before the old sink is abandoned.
+        with self._emit_lock:
+            if self._sink is None:
+                ctx = ModelInfraErrorContext.with_correlation(
+                    correlation_id=correlation_id,
+                    transport_type=EnumInfraTransportType.RUNTIME,
+                    operation="logging.emit",
+                    target_name="logging_handler",
+                )
+                raise RuntimeHostError(
+                    "Sink not initialized",
+                    context=ctx,
+                )
+
+            # Emit to sink (synchronous, non-blocking)
+            self._sink.emit(level, message, context)
+
+            # Capture metrics while still holding the lock for consistency
+            buffer_size = self._sink.buffer_size
+            drop_count = self._sink.drop_count
+
+        return ModelLoggingHandlerResponse(
+            status=EnumResponseStatus.SUCCESS,
+            operation="logging.emit",
+            message="Log entry buffered",
+            correlation_id=correlation_id,
+            buffer_size=buffer_size,
+            drop_count=drop_count,
+        )
+
+    def _handle_flush(
+        self,
+        correlation_id: UUID,
+    ) -> ModelLoggingHandlerResponse:
+        """Handle logging.flush operation.
+
+        Args:
+            correlation_id: Correlation ID for tracing
+
+        Returns:
+            ModelLoggingHandlerResponse with operation result
+        """
+        if self._sink is None:
+            ctx = ModelInfraErrorContext.with_correlation(
+                correlation_id=correlation_id,
+                transport_type=EnumInfraTransportType.RUNTIME,
+                operation="logging.flush",
+                target_name="logging_handler",
+            )
+            raise RuntimeHostError(
+                "Sink not initialized",
+                context=ctx,
+            )
+
+        # Capture pre-flush metrics
+        pre_buffer_size = self._sink.buffer_size
+
+        # Flush (synchronous I/O)
+        self._sink.flush()
+
+        return ModelLoggingHandlerResponse(
+            status=EnumResponseStatus.SUCCESS,
+            operation="logging.flush",
+            message=f"Flushed {pre_buffer_size} entries",
+            correlation_id=correlation_id,
+            buffer_size=self._sink.buffer_size,
+            drop_count=self._sink.drop_count,
+        )
+
+    async def _handle_configure(
+        self,
+        payload: dict[str, object],
+        correlation_id: UUID,
+    ) -> ModelLoggingHandlerResponse:
+        """Handle logging.configure operation.
+
+        Note: Configuration changes require re-creating the sink. To prevent
+        log loss, the swap-then-flush sequence ensures the atomic swap happens
+        first under _emit_lock, then the old sink is flushed after releasing
+        the lock.
+
+        Thread Safety:
+            The _emit_lock is acquired during the atomic swap operation only.
+            This ensures that:
+            1. All in-flight emits complete before we start the swap
+            2. No new emits can start during the swap
+            3. After swap completes, new emits immediately go to the new sink
+            4. Old sink is flushed after the lock is released (no blocking I/O under lock)
+
+        Reconfiguration Sequence:
+            1. Acquire _emit_lock
+            2. Swap to new sink atomically (under lock)
+            3. Update config reference (under lock)
+            4. Release _emit_lock
+            5. Flush old sink (outside lock - logs during flush go to new sink)
+            This order guarantees no log loss and minimizes lock hold time.
+
+        Args:
+            payload: Configuration payload (same fields as ModelLoggingHandlerConfig)
+            correlation_id: Correlation ID for tracing
+
+        Returns:
+            ModelLoggingHandlerResponse with operation result
+        """
+        async with self._config_lock:
+            if self._sink is None or self._config is None:
+                ctx = ModelInfraErrorContext.with_correlation(
+                    correlation_id=correlation_id,
+                    transport_type=EnumInfraTransportType.RUNTIME,
+                    operation="logging.configure",
+                    target_name="logging_handler",
+                )
+                raise RuntimeHostError(
+                    "Handler not initialized",
+                    context=ctx,
+                )
+
+            # Merge existing config with new values
+            current_dict = self._config.model_dump()
+            current_dict.update(payload)
+
+            # Validate new configuration BEFORE acquiring emit lock
+            # NOTE: Broad Exception catch is intentional here because Pydantic can raise
+            # various exception types (ValidationError, TypeError, ValueError) depending
+            # on the validation failure. We wrap all in ProtocolConfigurationError.
+            try:
+                new_config = ModelLoggingHandlerConfig.model_validate(current_dict)
+            except Exception as e:
+                ctx = ModelInfraErrorContext.with_correlation(
+                    correlation_id=correlation_id,
+                    transport_type=EnumInfraTransportType.RUNTIME,
+                    operation="logging.configure",
+                    target_name="logging_handler",
+                )
+                raise ProtocolConfigurationError(
+                    f"Invalid configuration: {e}",
+                    context=ctx,
+                ) from e
+
+            # Create new sink with updated config BEFORE acquiring emit lock
+            # to minimize time spent blocking emit operations.
+            # All config fields are applied including drop_policy.
+            new_sink = SinkLoggingStructured(
+                max_buffer_size=new_config.buffer_size,
+                output_format=new_config.output_format,
+                drop_policy=new_config.drop_policy,
+            )
+
+            # Critical section: acquire emit lock for atomic swap-then-flush.
+            # This prevents the race condition where:
+            # 1. emit reads self._sink -> gets old_sink
+            # 2. configure swaps to new_sink
+            # 3. emit calls old_sink.emit() -> log LOST (old sink abandoned)
+            #
+            # By swapping BEFORE flush and flushing AFTER releasing the lock:
+            # - New emits immediately go to the new sink (no blocking during flush)
+            # - Old sink's buffer is preserved until explicitly flushed
+            # - Lock hold time is minimized (fast swap, no I/O under lock)
+            with self._emit_lock:
+                # Step 1: Capture old sink reference and swap atomically
+                old_sink = self._sink
+                self._sink = new_sink
+                self._config = new_config
+
+                # Step 2: Reset drop count tracker for new sink
+                self._last_drop_count = 0
+
+            # Step 3: Flush old sink AFTER releasing lock.
+            # Logs emitted during this flush go to the new sink (not lost).
+            # This is safe because old_sink's buffer is independent.
+            if old_sink is not None:
+                old_sink.flush()
+
+            # Handle flush interval changes (outside emit lock - these are async)
+            if self._config.flush_interval_seconds > 0:
+                # Restart flush task if not running or interval changed
+                if self._flush_task is None or self._flush_task.done():
+                    self._flush_task = asyncio.create_task(
+                        self._periodic_flush_loop(),
+                        name="logging-handler-flush",
+                    )
+            # Cancel flush task if interval is 0
+            elif self._flush_task is not None and not self._flush_task.done():
+                self._flush_task.cancel()
+                try:
+                    await self._flush_task
+                except asyncio.CancelledError:
+                    pass
+                self._flush_task = None
+
+        return ModelLoggingHandlerResponse(
+            status=EnumResponseStatus.SUCCESS,
+            operation="logging.configure",
+            message="Configuration updated",
+            correlation_id=correlation_id,
+            buffer_size=self._sink.buffer_size if self._sink else 0,
+            drop_count=self._sink.drop_count if self._sink else 0,
+        )
+
+    def _parse_log_level(
+        self,
+        level_raw: object,
+        correlation_id: UUID,
+    ) -> EnumLogLevel:
+        """Parse log level from payload.
+
+        Accepts EnumLogLevel enum values or string level names.
+
+        Args:
+            level_raw: Raw level value from payload
+            correlation_id: Correlation ID for error context
+
+        Returns:
+            Parsed EnumLogLevel
+
+        Raises:
+            RuntimeHostError: If level is invalid
+        """
+        # Lazy import to avoid circular dependency
+        from omnibase_core.enums import EnumLogLevel
+
+        if isinstance(level_raw, EnumLogLevel):
+            return level_raw
+
+        if isinstance(level_raw, str):
+            # Try to match by value (case-insensitive)
+            level_upper = level_raw.upper()
+            for log_level in EnumLogLevel:
+                if log_level.value.upper() == level_upper:
+                    return log_level
+                if log_level.name.upper() == level_upper:
+                    return log_level
+
+        # Default to INFO for unknown levels
+        if level_raw is None:
+            return EnumLogLevel.INFO
+
+        ctx = ModelInfraErrorContext.with_correlation(
+            correlation_id=correlation_id,
+            transport_type=EnumInfraTransportType.RUNTIME,
+            operation="logging.emit",
+            target_name="logging_handler",
+        )
+        raise RuntimeHostError(
+            f"Invalid log level: {level_raw!r}. "
+            f"Valid levels: {[lvl.value for lvl in EnumLogLevel]}",
+            context=ctx,
+        )
+
+    async def _periodic_flush_loop(self) -> None:
+        """Background task for periodic buffer flush.
+
+        Runs until shutdown_event is set or task is cancelled.
+        Handles flush errors gracefully to prevent task crash.
+
+        Emits buffer utilization metrics before each flush:
+        - logging_buffer_utilization_ratio: Gauge showing buffer usage (0.0-1.0)
+        - logging_buffer_dropped_total: Counter of entries dropped since last flush
+        """
+        if self._shutdown_event is None:
+            return
+
+        while not self._shutdown_event.is_set():
+            try:
+                # Wait for flush interval or shutdown signal
+                flush_interval = (
+                    self._config.flush_interval_seconds
+                    if self._config is not None
+                    else 5.0
+                )
+                try:
+                    await asyncio.wait_for(
+                        self._shutdown_event.wait(),
+                        timeout=flush_interval,
+                    )
+                    # Shutdown signaled
+                    break
+                except TimeoutError:
+                    # Normal timeout - time to flush
+                    pass
+
+                # Perform flush with metrics emission
+                if self._sink is not None:
+                    # Emit buffer utilization metric (gauge: 0.0 to 1.0)
+                    # Safe from division by zero: max_buffer_size validated >= 1 at init
+                    max_size = self._sink.max_buffer_size
+                    current_size = self._sink.buffer_size
+                    utilization_ratio = current_size / max_size if max_size > 0 else 0.0
+
+                    logger.debug(
+                        "logging_buffer_utilization_ratio",
+                        extra={
+                            "handler": self.__class__.__name__,
+                            "metric_type": "gauge",
+                            "metric_name": "logging_buffer_utilization_ratio",
+                            "value": utilization_ratio,
+                            "buffer_size": current_size,
+                            "max_buffer_size": max_size,
+                        },
+                    )
+
+                    # Emit dropped logs counter (delta since last flush)
+                    current_drop_count = self._sink.drop_count
+                    dropped_since_last = current_drop_count - self._last_drop_count
+                    if dropped_since_last > 0:
+                        logger.warning(
+                            "logging_buffer_dropped_total",
+                            extra={
+                                "handler": self.__class__.__name__,
+                                "metric_type": "counter",
+                                "metric_name": "logging_buffer_dropped_total",
+                                "value": dropped_since_last,
+                                "total_dropped": current_drop_count,
+                            },
+                        )
+                    self._last_drop_count = current_drop_count
+
+                    # Perform flush
+                    self._sink.flush()
+
+            except asyncio.CancelledError:
+                # Task cancelled - exit gracefully
+                break
+            except Exception as e:
+                # NOTE: Broad Exception catch is intentional here to keep the
+                # background flush loop running despite transient I/O errors.
+                # This ensures the handler remains operational even if individual
+                # flush operations fail temporarily. The sleep prevents tight loops.
+                logger.warning(
+                    "Error in periodic flush loop: %s",
+                    e,
+                    extra={
+                        "handler": self.__class__.__name__,
+                        "error_type": type(e).__name__,
+                    },
+                )
+                # Brief sleep to prevent tight error loop on persistent failures
+                await asyncio.sleep(1.0)
+
+    def describe(self) -> dict[str, object]:
+        """Return handler metadata and capabilities for introspection.
+
+        This method exposes the handler's type classification along with its
+        operational configuration and capabilities.
+
+        Returns:
+            dict containing:
+                - handler_type: Architectural role (infra_handler)
+                - handler_category: Behavioral classification (effect)
+                - supported_operations: List of supported operations
+                - initialized: Whether the handler is initialized
+                - version: Handler version string
+                - config: Current configuration (if initialized)
+                - buffer_size: Current buffer size (if initialized)
+                - max_buffer_size: Maximum buffer capacity (if initialized)
+                - drop_count: Total dropped entries (if initialized)
+                - buffer_utilization_ratio: Current utilization (0.0-1.0)
+
+        See Also:
+            - handler_type property: Full documentation of architectural role
+            - handler_category property: Full documentation of behavioral classification
+        """
+        result: dict[str, object] = {
+            "handler_type": self.handler_type.value,
+            "handler_category": self.handler_category.value,
+            "supported_operations": sorted(SUPPORTED_OPERATIONS),
+            "initialized": self._initialized,
+            "version": "0.1.0-mvp",
+        }
+
+        if self._initialized and self._config is not None:
+            result["config"] = {
+                "buffer_size": self._config.buffer_size,
+                "flush_interval_seconds": self._config.flush_interval_seconds,
+                "output_format": self._config.output_format,
+                "drop_policy": self._config.drop_policy,
+            }
+
+        if self._sink is not None:
+            result["buffer_size"] = self._sink.buffer_size
+            result["max_buffer_size"] = self._sink.max_buffer_size
+            result["drop_count"] = self._sink.drop_count
+            # Include utilization ratio for monitoring
+            max_size = self._sink.max_buffer_size
+            result["buffer_utilization_ratio"] = (
+                self._sink.buffer_size / max_size if max_size > 0 else 0.0
+            )
+
+        return result
+
+
+__all__: list[str] = [
+    "HandlerLoggingStructured",
+    "SUPPORTED_OPERATIONS",
+]