omnibase_infra 0.2.6__py3-none-any.whl

omnibase_infra/handlers/handler_http.py

@@ -0,0 +1,926 @@
+# SPDX-License-Identifier: MIT
+# Copyright (c) 2025 OmniNode Team
+"""HTTP REST Handler - MVP implementation using httpx async client.
+
+Supports GET and POST operations with 30-second fixed timeout.
+PUT, DELETE, PATCH deferred to Beta. Retry logic and rate limiting deferred to Beta.
+
+Note:
+    Environment variable configuration (ONEX_HTTP_TIMEOUT, ONEX_HTTP_MAX_REQUEST_SIZE,
+    ONEX_HTTP_MAX_RESPONSE_SIZE) is parsed at module import time, not at handler
+    instantiation. This means:
+
+    - Changes to environment variables require application restart to take effect
+    - Tests should use ``unittest.mock.patch.dict(os.environ, ...)`` before importing,
+      or use ``importlib.reload()`` to re-import the module after patching
+    - This is an intentional design choice for startup-time validation
+"""
+
+from __future__ import annotations
+
+import json
+import logging
+from uuid import UUID, uuid4
+
+import httpx
+
+from omnibase_core.container import ModelONEXContainer
+from omnibase_core.models.dispatch import ModelHandlerOutput
+from omnibase_infra.enums import (
+    EnumHandlerType,
+    EnumHandlerTypeCategory,
+    EnumInfraTransportType,
+)
+from omnibase_infra.errors import (
+    InfraConnectionError,
+    InfraTimeoutError,
+    InfraUnavailableError,
+    ModelInfraErrorContext,
+    ModelTimeoutErrorContext,
+    ProtocolConfigurationError,
+    RuntimeHostError,
+)
+from omnibase_infra.handlers.models.http import ModelHttpBodyContent
+from omnibase_infra.mixins import MixinEnvelopeExtraction
+from omnibase_infra.utils import parse_env_float, parse_env_int
+
+logger = logging.getLogger(__name__)
+
+
+_DEFAULT_TIMEOUT_SECONDS: float = parse_env_float(
+    "ONEX_HTTP_TIMEOUT",
+    30.0,
+    min_value=1.0,
+    max_value=300.0,
+    transport_type=EnumInfraTransportType.HTTP,
+    service_name="http_handler",
+)
+_DEFAULT_MAX_REQUEST_SIZE: int = parse_env_int(
+    "ONEX_HTTP_MAX_REQUEST_SIZE",
+    10 * 1024 * 1024,
+    min_value=1024,
+    max_value=104857600,
+    transport_type=EnumInfraTransportType.HTTP,
+    service_name="http_handler",
+)  # 10 MB default, min 1 KB, max 100 MB
+_DEFAULT_MAX_RESPONSE_SIZE: int = parse_env_int(
+    "ONEX_HTTP_MAX_RESPONSE_SIZE",
+    50 * 1024 * 1024,
+    min_value=1024,
+    max_value=104857600,
+    transport_type=EnumInfraTransportType.HTTP,
+    service_name="http_handler",
+)  # 50 MB default, min 1 KB, max 100 MB
+_SUPPORTED_OPERATIONS: frozenset[str] = frozenset({"http.get", "http.post"})
+# Streaming chunk size for responses without Content-Length header
+_STREAMING_CHUNK_SIZE: int = 8192  # 8 KB chunks
+
+# Handler ID for ModelHandlerOutput
+HANDLER_ID_HTTP: str = "http-handler"
+
+# Size category thresholds for sanitized logging
+_SIZE_THRESHOLD_KB: int = 1024  # 1 KB
+_SIZE_THRESHOLD_MB: int = 1024 * 1024  # 1 MB
+_SIZE_THRESHOLD_10MB: int = 10 * 1024 * 1024  # 10 MB
+
+
+def _categorize_size(size: int) -> str:
+    """Categorize byte size into security-safe categories.
+
+    This prevents exact payload sizes from being exposed in error messages
+    and logs, which could help attackers probe size limits.
+
+    Args:
+        size: Size in bytes
+
+    Returns:
+        Size category: "small", "medium", "large", or "very_large"
+    """
+    if size < _SIZE_THRESHOLD_KB:
+        return "small"
+    elif size < _SIZE_THRESHOLD_MB:
+        return "medium"
+    elif size < _SIZE_THRESHOLD_10MB:
+        return "large"
+    else:
+        return "very_large"
+
+
+class HandlerHttpRest(MixinEnvelopeExtraction):
+    """HTTP REST protocol handler using httpx async client (MVP: GET, POST only).
+
+    Security Features:
+        - Configurable request/response size limits to prevent DoS attacks
+        - Pre-read Content-Length validation to prevent memory exhaustion
+        - Streaming body validation for chunked transfer encoding
+    """
+
+    def __init__(self, container: ModelONEXContainer) -> None:
+        """Initialize HandlerHttpRest with ONEX container for dependency injection.
+
+        Args:
+            container: ONEX container for dependency injection.
+        """
+        self._container = container
+        self._client: httpx.AsyncClient | None = None
+        self._timeout: float = _DEFAULT_TIMEOUT_SECONDS
+        self._max_request_size: int = _DEFAULT_MAX_REQUEST_SIZE
+        self._max_response_size: int = _DEFAULT_MAX_RESPONSE_SIZE
+        self._initialized: bool = False
+
+    @property
+    def handler_type(self) -> EnumHandlerType:
+        """Return the architectural role of this handler.
+
+        Returns:
+            EnumHandlerType.INFRA_HANDLER - This handler is an infrastructure
+            protocol/transport handler (as opposed to NODE_HANDLER for event
+            processing, PROJECTION_HANDLER for read models, or COMPUTE_HANDLER
+            for 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)
+            - transport_type: Specific transport protocol (HTTP/DATABASE/etc.)
+            - 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 (external HTTP requests). 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, service calls)
+            - NONDETERMINISTIC_COMPUTE: Pure but not deterministic (UUID, random)
+
+        See Also:
+            - handler_type: Architectural role (INFRA_HANDLER/NODE_HANDLER/etc.)
+            - transport_type: Specific transport protocol (HTTP/DATABASE/etc.)
+            - docs/architecture/HANDLER_PROTOCOL_DRIVEN_ARCHITECTURE.md
+        """
+        return EnumHandlerTypeCategory.EFFECT
+
+    @property
+    def transport_type(self) -> EnumInfraTransportType:
+        """Return the transport protocol identifier for this handler.
+
+        Returns:
+            EnumInfraTransportType.HTTP - This handler uses HTTP/REST protocol.
+
+        Note:
+            transport_type identifies the specific transport/protocol this handler
+            uses. It is the third dimension of the handler type system, alongside
+            handler_type (architectural role) and handler_category (behavioral
+            classification).
+
+            The three dimensions together form a complete handler classification:
+            - handler_type: INFRA_HANDLER (what it is architecturally)
+            - handler_category: EFFECT (how it behaves at runtime)
+            - transport_type: HTTP (what protocol it uses)
+
+        See Also:
+            - handler_type: Architectural role
+            - handler_category: Behavioral classification
+            - docs/architecture/HANDLER_PROTOCOL_DRIVEN_ARCHITECTURE.md
+        """
+        return EnumInfraTransportType.HTTP
+
+    async def initialize(self, config: dict[str, object]) -> None:
+        """Initialize HTTP client with configurable timeout and size limits.
+
+        Args:
+            config: Configuration dict containing:
+                - max_request_size: Optional max request body size in bytes (default: 10 MB)
+                - max_response_size: Optional max response body size in bytes (default: 50 MB)
+                - correlation_id: Optional UUID or string for error tracing
+
+        Raises:
+            ProtocolConfigurationError: If client initialization fails.
+
+        Security:
+            Size limits provide DoS (Denial of Service) protection by preventing:
+            - Memory exhaustion from oversized request bodies
+            - Memory exhaustion from malicious/misconfigured server responses
+            - Resource starvation attacks via large payload processing
+
+            Content-Length headers are validated BEFORE reading response bodies,
+            and streaming validation protects against chunked transfer encoding attacks.
+
+        Note:
+            Size limit violations raise InfraUnavailableError with sanitized size
+            categories (small/medium/large/very_large) - exact sizes are not exposed
+            in error messages to prevent attackers from probing limits.
+        """
+        # Generate correlation_id for initialization tracing
+        init_correlation_id = uuid4()
+
+        logger.info(
+            "Initializing %s",
+            self.__class__.__name__,
+            extra={
+                "handler": self.__class__.__name__,
+                "correlation_id": str(init_correlation_id),
+            },
+        )
+
+        try:
+            self._timeout = _DEFAULT_TIMEOUT_SECONDS
+
+            # Extract configurable size limits
+            max_request_raw = config.get("max_request_size")
+            if max_request_raw is not None:
+                if isinstance(max_request_raw, int) and max_request_raw > 0:
+                    self._max_request_size = max_request_raw
+                else:
+                    logger.warning(
+                        "Invalid max_request_size config value ignored, using default",
+                        extra={
+                            "provided_value": max_request_raw,
+                            "default_value": self._max_request_size,
+                        },
+                    )
+
+            max_response_raw = config.get("max_response_size")
+            if max_response_raw is not None:
+                if isinstance(max_response_raw, int) and max_response_raw > 0:
+                    self._max_response_size = max_response_raw
+                else:
+                    logger.warning(
+                        "Invalid max_response_size config value ignored, using default",
+                        extra={
+                            "provided_value": max_response_raw,
+                            "default_value": self._max_response_size,
+                        },
+                    )
+
+            self._client = httpx.AsyncClient(
+                timeout=httpx.Timeout(self._timeout),
+                follow_redirects=True,
+            )
+            self._initialized = True
+            logger.info(
+                "%s initialized successfully",
+                self.__class__.__name__,
+                extra={
+                    "handler": self.__class__.__name__,
+                    "timeout_seconds": self._timeout,
+                    "max_request_size_bytes": self._max_request_size,
+                    "max_response_size_bytes": self._max_response_size,
+                    "correlation_id": str(init_correlation_id),
+                },
+            )
+        except Exception as e:
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation="initialize",
+                target_name="http_handler",
+                correlation_id=init_correlation_id,
+            )
+            raise ProtocolConfigurationError(
+                "Failed to initialize HTTP handler", context=ctx
+            ) from e
+
+    async def shutdown(self) -> None:
+        """Close HTTP client and release resources."""
+        if self._client is not None:
+            await self._client.aclose()
+            self._client = None
+        self._initialized = False
+        logger.info("HandlerHttpRest shutdown complete")
+
+    async def execute(
+        self, envelope: dict[str, object]
+    ) -> ModelHandlerOutput[dict[str, object]]:
+        """Execute HTTP operation (http.get or http.post) from envelope.
+
+        Args:
+            envelope: Request envelope containing:
+                - operation: "http.get" or "http.post"
+                - payload: dict with "url" (required), "headers" (optional), "body" (optional for POST)
+                - correlation_id: Optional correlation ID for tracing
+                - envelope_id: Optional envelope ID for causality tracking
+
+        Returns:
+            ModelHandlerOutput[dict[str, object]] containing:
+                - result: dict with status, payload (status_code, headers, body), and correlation_id
+                - input_envelope_id: UUID for causality tracking
+                - correlation_id: UUID for request/response correlation
+                - handler_id: "http-handler"
+        """
+        correlation_id = self._extract_correlation_id(envelope)
+        input_envelope_id = self._extract_envelope_id(envelope)
+
+        if not self._initialized or self._client is None:
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation="execute",
+                target_name="http_handler",
+                correlation_id=correlation_id,
+            )
+            raise RuntimeHostError(
+                "HandlerHttpRest not initialized. Call initialize() first.", context=ctx
+            )
+
+        operation = envelope.get("operation")
+        if not isinstance(operation, str):
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation="execute",
+                target_name="http_handler",
+                correlation_id=correlation_id,
+            )
+            raise ProtocolConfigurationError(
+                "Missing or invalid 'operation' in envelope", context=ctx
+            )
+
+        if operation not in _SUPPORTED_OPERATIONS:
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation=operation,
+                target_name="http_handler",
+                correlation_id=correlation_id,
+            )
+            raise ProtocolConfigurationError(
+                f"Operation '{operation}' not supported in MVP. Available: {', '.join(sorted(_SUPPORTED_OPERATIONS))}",
+                context=ctx,
+            )
+
+        payload = envelope.get("payload")
+        if not isinstance(payload, dict):
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation=operation,
+                target_name="http_handler",
+                correlation_id=correlation_id,
+            )
+            raise ProtocolConfigurationError(
+                "Missing or invalid 'payload' in envelope", context=ctx
+            )
+
+        url = payload.get("url")
+        if not isinstance(url, str) or not url:
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation=operation,
+                target_name="http_handler",
+                correlation_id=correlation_id,
+            )
+            raise ProtocolConfigurationError(
+                "Missing or invalid 'url' in payload", context=ctx
+            )
+
+        headers = self._extract_headers(payload, operation, url, correlation_id)
+
+        if operation == "http.get":
+            return await self._execute_request(
+                "GET", url, headers, None, correlation_id, input_envelope_id, None
+            )
+        else:  # http.post
+            body = payload.get("body")
+            # Validate request body size and get pre-serialized bytes for dict bodies.
+            # This avoids double serialization - dict bodies are serialized once here
+            # and the cached bytes are passed to _execute_request().
+            pre_serialized = self._validate_request_size(body, correlation_id)
+            return await self._execute_request(
+                "POST",
+                url,
+                headers,
+                body,
+                correlation_id,
+                input_envelope_id,
+                pre_serialized,
+            )
+
+    def _extract_headers(
+        self,
+        payload: dict[str, object],
+        operation: str,
+        url: str,
+        correlation_id: UUID,
+    ) -> dict[str, str]:
+        """Extract and validate headers from payload."""
+        headers_raw = payload.get("headers")
+        if headers_raw is None:
+            return {}
+        if isinstance(headers_raw, dict):
+            return {str(k): str(v) for k, v in headers_raw.items()}
+        ctx = ModelInfraErrorContext(
+            transport_type=EnumInfraTransportType.HTTP,
+            operation=operation,
+            target_name=url,
+            correlation_id=correlation_id,
+        )
+        raise ProtocolConfigurationError(
+            "Invalid 'headers' in payload - must be a dict", context=ctx
+        )
+
+    def _validate_request_size(
+        self, body: object, correlation_id: UUID
+    ) -> bytes | None:
+        """Validate request body size and cache serialized bytes for dict bodies.
+
+        For dict bodies, this method serializes the body once and returns the
+        serialized bytes. This avoids double serialization - the returned bytes
+        can be passed directly to _execute_request() instead of re-serializing.
+
+        Args:
+            body: Request body (str, dict, bytes, or None)
+            correlation_id: Correlation ID for error context
+
+        Returns:
+            For dict bodies: The pre-serialized JSON bytes (cached for reuse)
+            For other body types: None (no caching needed)
+
+        Raises:
+            InfraUnavailableError: If body size exceeds max_request_size limit.
+        """
+        if body is None:
+            return None
+
+        size: int = 0
+        serialized_bytes: bytes | None = None
+
+        if isinstance(body, str):
+            size = len(body.encode("utf-8"))
+        elif isinstance(body, dict):
+            # DESIGN TRADEOFF: Double-Serialization Avoidance
+            #
+            # We serialize dict bodies once here during validation and cache the bytes.
+            # The cached bytes are then passed to _execute_request() via the return value,
+            # avoiding re-serialization when building the HTTP request.
+            #
+            # Tradeoff considerations:
+            # - Memory: Serialized bytes are held in memory during validation and request
+            #   execution. For large payloads near the size limit, this adds ~10MB overhead.
+            # - Performance: Single serialization is faster than serializing twice (once
+            #   for size check, once for request body).
+            # - Alternative: We could serialize twice (once here for size, once in execute)
+            #   which would use less peak memory but double the CPU cost for serialization.
+            #
+            # Current approach prioritizes CPU efficiency over peak memory usage, which is
+            # appropriate since we enforce max_request_size limits anyway.
+            try:
+                serialized_bytes = json.dumps(body).encode("utf-8")
+                size = len(serialized_bytes)
+            except (TypeError, ValueError):
+                # If we can't serialize, skip validation and let execute() handle it
+                return None
+        elif isinstance(body, bytes):
+            size = len(body)
+        else:
+            # Unknown type - skip validation, let execute() handle serialization
+            return None
+
+        if size > self._max_request_size:
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation="validate_request_size",
+                target_name="http_handler",
+                correlation_id=correlation_id,
+            )
+            raise InfraUnavailableError(
+                f"Request body size ({_categorize_size(size)}) exceeds configured limit",
+                context=ctx,
+            )
+
+        logger.debug(
+            "Request size validated",
+            extra={
+                "request_size_category": _categorize_size(size),
+                "limit": self._max_request_size,
+                "correlation_id": str(correlation_id),
+            },
+        )
+
+        # Return cached serialized bytes for dict bodies, None for other types
+        return serialized_bytes
+
+    def _validate_content_length_header(
+        self, response: httpx.Response, url: str, correlation_id: UUID
+    ) -> None:
+        """Validate Content-Length header BEFORE reading response body.
+
+        This prevents memory exhaustion by rejecting large responses before
+        they are loaded into memory. This is critical for security as it
+        prevents denial-of-service attacks via large response payloads.
+
+        Args:
+            response: The httpx Response object (body not yet read)
+            url: Target URL for error context
+            correlation_id: Correlation ID for error context
+
+        Raises:
+            InfraUnavailableError: If Content-Length exceeds max_response_size limit.
+        """
+        content_length_header = response.headers.get("content-length")
+        if content_length_header is None:
+            # No Content-Length header - will use streaming validation
+            return
+
+        try:
+            content_length = int(content_length_header)
+        except ValueError:
+            # Invalid Content-Length header - log warning and proceed with body-based validation
+            logger.warning(
+                "Invalid Content-Length header value",
+                extra={
+                    "content_length_header": content_length_header,
+                    "url": url,
+                    "correlation_id": str(correlation_id),
+                },
+            )
+            return
+
+        if content_length < 0:
+            # Negative Content-Length is invalid per HTTP spec - log warning and proceed
+            # with body-based validation (streaming protection still applies)
+            logger.warning(
+                "Negative Content-Length header value ignored",
+                extra={
+                    "content_length_header": content_length_header,
+                    "url": url,
+                    "correlation_id": str(correlation_id),
+                },
+            )
+            return
+
+        if content_length > self._max_response_size:
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation="validate_content_length",
+                target_name=url,
+                correlation_id=correlation_id,
+            )
+            raise InfraUnavailableError(
+                f"Response Content-Length ({_categorize_size(content_length)}) exceeds configured limit",
+                context=ctx,
+            )
+
+        logger.debug(
+            "Content-Length header validated",
+            extra={
+                "content_length_category": _categorize_size(content_length),
+                "limit": self._max_response_size,
+                "url": url,
+                "correlation_id": str(correlation_id),
+            },
+        )
+
+    async def _read_response_body_with_limit(
+        self, response: httpx.Response, url: str, correlation_id: UUID
+    ) -> bytes:
+        """Read response body with streaming size limit enforcement.
+
+        For responses without Content-Length header (e.g., chunked transfer encoding),
+        this method reads the body in chunks and tracks the total size, stopping
+        and raising an error if the limit is exceeded.
+
+        Security Note: This method provides DoS protection for chunked transfer
+        encoding responses where Content-Length is not available for pre-read
+        validation. By enforcing limits during streaming, we prevent memory
+        exhaustion attacks from maliciously large chunked responses.
+
+        Args:
+            response: The httpx Response object
+            url: Target URL for error context
+            correlation_id: Correlation ID for error context
+
+        Returns:
+            The complete response body as bytes
+
+        Raises:
+            InfraUnavailableError: If body size exceeds max_response_size during streaming.
+        """
+        chunks: list[bytes] = []
+        total_size: int = 0
+
+        async for chunk in response.aiter_bytes(chunk_size=_STREAMING_CHUNK_SIZE):
+            total_size += len(chunk)
+            if total_size > self._max_response_size:
+                logger.warning(
+                    "Response body exceeded size limit during streaming read",
+                    extra={
+                        "size_category": _categorize_size(total_size),
+                        "limit": self._max_response_size,
+                        "url": url,
+                        "correlation_id": str(correlation_id),
+                    },
+                )
+                ctx = ModelInfraErrorContext(
+                    transport_type=EnumInfraTransportType.HTTP,
+                    operation="read_response_body",
+                    target_name=url,
+                    correlation_id=correlation_id,
+                )
+                raise InfraUnavailableError(
+                    f"Response body size ({_categorize_size(total_size)}) exceeds configured limit during streaming read",
+                    context=ctx,
+                )
+            chunks.append(chunk)
+
+        return b"".join(chunks)
+
+    def _prepare_request_content(
+        self,
+        method: str,
+        headers: dict[str, str],
+        body_content: ModelHttpBodyContent,
+        ctx: ModelInfraErrorContext,
+    ) -> tuple[bytes | str | None, dict[str, object] | None, dict[str, str]]:
+        """Prepare request content for HTTP request.
+
+        Handles body serialization for POST requests, managing pre-serialized bytes
+        from size validation and various body types (dict, str, other JSON-serializable).
+
+        Args:
+            method: HTTP method (GET, POST)
+            headers: Request headers (will be copied, not mutated)
+            body_content: Model containing body and optional pre-serialized bytes
+            ctx: Error context for exceptions
+
+        Returns:
+            Tuple of (request_content, request_json, request_headers):
+                - request_content: bytes or str content for the request
+                - request_json: dict for httpx json= parameter (mutually exclusive with content)
+                - request_headers: Headers dict (possibly with Content-Type added)
+
+        Raises:
+            ProtocolConfigurationError: If body is not JSON-serializable.
+        """
+        request_content: bytes | str | None = None
+        request_json: dict[str, object] | None = None
+        request_headers = dict(headers)  # Copy to avoid mutating caller's headers
+
+        body = body_content.body
+        pre_serialized = body_content.pre_serialized
+
+        if method != "POST" or body is None:
+            return request_content, request_json, request_headers
+
+        if pre_serialized is not None:
+            # Use pre-serialized bytes from _validate_request_size to avoid
+            # double serialization. Set Content-Type header since we're using
+            # content= instead of json= parameter.
+            request_content = pre_serialized
+            if "content-type" not in {k.lower() for k in request_headers}:
+                request_headers["Content-Type"] = "application/json"
+        elif isinstance(body, dict):
+            # Fallback for dict bodies without pre-serialized content
+            # (shouldn't happen in normal flow, but handles edge cases)
+            request_json = body
+        elif isinstance(body, str):
+            request_content = body
+        else:
+            try:
+                request_content = json.dumps(body)
+            except TypeError as e:
+                raise ProtocolConfigurationError(
+                    f"Body is not JSON-serializable: {type(body).__name__}",
+                    context=ctx,
+                ) from e
+
+        return request_content, request_json, request_headers
+
+    async def _execute_request(
+        self,
+        method: str,
+        url: str,
+        headers: dict[str, str],
+        body: object,
+        correlation_id: UUID,
+        input_envelope_id: UUID,
+        pre_serialized: bytes | None = None,
+    ) -> ModelHandlerOutput[dict[str, object]]:
+        """Execute HTTP request with pre-read response size validation.
+
+        Uses httpx streaming to validate Content-Length header BEFORE reading
+        the response body into memory, preventing memory exhaustion attacks.
+
+        Args:
+            method: HTTP method (GET, POST)
+            url: Target URL
+            headers: Request headers
+            body: Request body (used only if pre_serialized is None)
+            correlation_id: Correlation ID for tracing
+            input_envelope_id: Envelope ID for causality tracking
+            pre_serialized: Pre-serialized JSON bytes for dict bodies (from
+                _validate_request_size). When provided, this is used directly
+                instead of re-serializing the body, avoiding double serialization.
+
+        Returns:
+            ModelHandlerOutput[dict[str, object]] with wrapped response data
+        """
+        if self._client is None:
+            ctx = ModelInfraErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation=f"http.{method.lower()}",
+                target_name=url,
+                correlation_id=correlation_id,
+            )
+            raise RuntimeHostError(
+                "HandlerHttpRest not initialized - call initialize() first", context=ctx
+            )
+
+        ctx = ModelInfraErrorContext(
+            transport_type=EnumInfraTransportType.HTTP,
+            operation=f"http.{method.lower()}",
+            target_name=url,
+            correlation_id=correlation_id,
+        )
+
+        # Prepare request content for POST
+        body_content = ModelHttpBodyContent(body=body, pre_serialized=pre_serialized)
+        request_content, request_json, request_headers = self._prepare_request_content(
+            method, headers, body_content, ctx
+        )
+
+        try:
+            # Use streaming request to get response headers before reading body
+            # This allows us to check Content-Length before loading body into memory
+            async with self._client.stream(
+                method,
+                url,
+                headers=request_headers,
+                content=request_content,
+                json=request_json,
+            ) as response:
+                # CRITICAL: Validate Content-Length header BEFORE reading body
+                # This prevents memory exhaustion from large responses
+                self._validate_content_length_header(response, url, correlation_id)
+
+                # Read body with streaming size limit enforcement
+                # For responses without Content-Length, this stops early if limit exceeded
+                response_body_bytes = await self._read_response_body_with_limit(
+                    response, url, correlation_id
+                )
+
+                return self._build_response_from_bytes(
+                    response, response_body_bytes, correlation_id, input_envelope_id
+                )
+
+        except httpx.TimeoutException as e:
+            timeout_ctx = ModelTimeoutErrorContext(
+                transport_type=EnumInfraTransportType.HTTP,
+                operation=f"http.{method.lower()}",
+                target_name=url,
+                correlation_id=correlation_id,
+                timeout_seconds=self._timeout,
+            )
+            raise InfraTimeoutError(
+                f"HTTP {method} request timed out after {self._timeout}s",
+                context=timeout_ctx,
+            ) from e
+        except httpx.ConnectError as e:
+            raise InfraConnectionError(
+                f"Failed to connect to {url}", context=ctx
+            ) from e
+        except httpx.HTTPError as e:
+            raise InfraConnectionError(
+                f"HTTP error during {method} request: {type(e).__name__}", context=ctx
+            ) from e
+
+    def _build_response_from_bytes(
+        self,
+        response: httpx.Response,
+        body_bytes: bytes,
+        correlation_id: UUID,
+        input_envelope_id: UUID,
+    ) -> ModelHandlerOutput[dict[str, object]]:
+        """Build response envelope from httpx Response and pre-read body bytes.
+
+        This method is used with streaming responses where the body has already
+        been read with size limit enforcement.
+
+        Args:
+            response: The httpx Response object (headers already available)
+            body_bytes: The pre-read response body bytes
+            correlation_id: Correlation ID for tracing
+            input_envelope_id: Envelope ID for causality tracking
+
+        Returns:
+            ModelHandlerOutput wrapping response dict with status, payload, and correlation_id
+        """
+        content_type = response.headers.get("content-type", "")
+        body: object
+
+        # TODO(OMN-43): When rate limiting is implemented, extract and log rate limit
+        # response headers: x-ratelimit-remaining, x-ratelimit-limit, x-ratelimit-reset
+        # These headers will be added to the debug log metadata below for observability.
+
+        logger.debug(
+            "Response body received",
+            extra={
+                "body_size": len(body_bytes),
+                "size_utilization_pct": round(
+                    (len(body_bytes) / self._max_response_size) * 100, 2
+                ),
+                "content_type": content_type,
+                "status_code": response.status_code,
+                "correlation_id": str(correlation_id),
+                # TODO(OMN-43): Add rate limit metadata here when rate limiting is implemented:
+                # "ratelimit_remaining": response.headers.get("x-ratelimit-remaining"),
+                # "ratelimit_limit": response.headers.get("x-ratelimit-limit"),
+                # "ratelimit_reset": response.headers.get("x-ratelimit-reset"),
+            },
+        )
+
+        # Decode bytes to string first
+        try:
+            body_text = body_bytes.decode("utf-8")
+        except UnicodeDecodeError:
+            # If UTF-8 decoding fails, try latin-1 as fallback
+            body_text = body_bytes.decode("latin-1")
+
+        # Try to parse as JSON if content type indicates JSON
+        if "application/json" in content_type:
+            try:
+                body = json.loads(body_text)
+            except json.JSONDecodeError:
+                body = body_text
+        else:
+            body = body_text
+
+        return ModelHandlerOutput.for_compute(
+            input_envelope_id=input_envelope_id,
+            correlation_id=correlation_id,
+            handler_id=HANDLER_ID_HTTP,
+            result={
+                "status": "success",
+                "payload": {
+                    "status_code": response.status_code,
+                    "headers": dict(response.headers),
+                    "body": body,
+                },
+                "correlation_id": str(correlation_id),
+            },
+        )
+
+    def describe(self) -> dict[str, object]:
+        """Return handler metadata and capabilities for introspection.
+
+        This method exposes the handler's three-dimensional type classification
+        along with its operational configuration and capabilities.
+
+        Returns:
+            dict containing:
+                - handler_type: Architectural role from handler_type property
+                  (e.g., "infra_handler"). See EnumHandlerType for valid values.
+                - handler_category: Behavioral classification from handler_category
+                  property (e.g., "effect"). See EnumHandlerTypeCategory for valid values.
+                - transport_type: Protocol identifier from transport_type property
+                  (e.g., "http"). See EnumInfraTransportType for valid values.
+                - supported_operations: List of supported operations
+                - timeout_seconds: Request timeout in seconds
+                - max_request_size: Maximum request body size in bytes
+                - max_response_size: Maximum response body size in bytes
+                - initialized: Whether the handler is initialized
+                - version: Handler version string
+
+        Note:
+            The handler_type, handler_category, and transport_type fields form the
+            three-dimensional handler classification system:
+
+            1. handler_type (architectural role): Determines lifecycle and invocation
+               patterns. This handler is INFRA_HANDLER (protocol/transport handler).
+
+            2. handler_category (behavioral classification): Determines security rules
+               and replay safety. This handler is EFFECT (side-effecting I/O).
+
+            3. transport_type (protocol identifier): Identifies the specific transport.
+               This handler uses HTTP protocol.
+
+        See Also:
+            - handler_type property: Full documentation of architectural role
+            - handler_category property: Full documentation of behavioral classification
+            - transport_type property: Full documentation of transport identifier
+            - docs/architecture/HANDLER_PROTOCOL_DRIVEN_ARCHITECTURE.md
+        """
+        return {
+            "handler_type": self.handler_type.value,
+            "handler_category": self.handler_category.value,
+            "transport_type": self.transport_type.value,
+            "supported_operations": sorted(_SUPPORTED_OPERATIONS),
+            "timeout_seconds": self._timeout,
+            "max_request_size": self._max_request_size,
+            "max_response_size": self._max_response_size,
+            "initialized": self._initialized,
+            "version": "0.1.0-mvp",
+        }
+
+
+__all__: list[str] = ["HandlerHttpRest"]