openlit 1.33.8__py3-none-any.whl → 1.33.9__py3-none-any.whl

openlit/instrumentation/azure_ai_inference/async_azure_ai_inference.py

@@ -1,15 +1,20 @@
-# pylint: disable=duplicate-code, broad-exception-caught, too-many-statements, unused-argument, possibly-used-before-assignment, protected-access
 """
 Module for monitoring Azure AI Inference API calls.
 """
 
 import logging
+import time
 from opentelemetry.trace import SpanKind, Status, StatusCode
-from opentelemetry.sdk.resources import TELEMETRY_SDK_NAME
+from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT
 from openlit.__helpers import (
-    handle_exception,
     get_chat_model_cost,
     get_embed_model_cost,
+    handle_exception,
+    response_as_dict,
+    calculate_ttft,
+    calculate_tbt,
+    create_metrics_attributes,
+    set_server_address_and_port,
     general_tokens
 )
 from openlit.semcov import SemanticConvetion
@@ -17,13 +22,12 @@ from openlit.semcov import SemanticConvetion
 # Initialize logger for logging potential issues and operations
 logger = logging.getLogger(__name__)
 
-def async_complete(gen_ai_endpoint, version, environment, application_name,
+def async_complete(version, environment, application_name,
                      tracer, pricing_info, trace_content, metrics, disable_metrics):
     """
     Generates a telemetry wrapper for chat to collect metrics.
 
     Args:
-        gen_ai_endpoint: Endpoint identifier for logging and tracing.
         version: Version of the monitoring package.
         environment: Deployment environment (e.g., production, staging).
         application_name: Name of the application using the Azure AI Inference API.
@@ -35,165 +39,275 @@ def async_complete(gen_ai_endpoint, version, environment, application_name,
         A function that wraps the chat method to add telemetry.
     """
 
+    class TracedAsyncStream:
+        """
+        Wrapper for streaming responses to collect metrics and trace data.
+        Wraps the 'az.ai.inference.AsyncStream' response to collect message IDs and aggregated response.
+
+        This class implements the '__aiter__' and '__anext__' methods that
+        handle asynchronous streaming responses.
+
+        This class also implements '__aenter__' and '__aexit__' methods that
+        handle asynchronous context management protocol.
+        """
+        def __init__(
+                self,
+                wrapped,
+                span,
+                kwargs,
+                server_address,
+                server_port,
+                **args,
+            ):
+            self.__wrapped__ = wrapped
+            self._span = span
+            # Placeholder for aggregating streaming response
+            self._llmresponse = ""
+            self._response_id = ""
+            self._response_model = ""
+            self._finish_reason = ""
+            self._system_fingerprint = ""
+
+            self._args = args
+            self._kwargs = kwargs
+            self._start_time = time.time()
+            self._end_time = None
+            self._timestamps = []
+            self._ttft = 0
+            self._tbt = 0
+            self._server_address = server_address
+            self._server_port = server_port
+
+        async def __aenter__(self):
+            await self.__wrapped__.__aenter__()
+            return self
+
+        async def __aexit__(self, exc_type, exc_value, traceback):
+            await self.__wrapped__.__aexit__(exc_type, exc_value, traceback)
+
+        def __aiter__(self):
+            return self
+
+        async def __getattr__(self, name):
+            """Delegate attribute access to the wrapped object."""
+            return getattr(await self.__wrapped__, name)
+
+        async def __anext__(self):
+            try:
+                chunk = await self.__wrapped__.__anext__()
+                end_time = time.time()
+                # Record the timestamp for the current chunk
+                self._timestamps.append(end_time)
+
+                if len(self._timestamps) == 1:
+                    # Calculate time to first chunk
+                    self._ttft = calculate_ttft(self._timestamps, self._start_time)
+
+                chunked = response_as_dict(chunk)
+                # Collect message IDs and aggregated response from events
+                if (len(chunked.get('choices')) > 0 and ('delta' in chunked.get('choices')[0] and
+                    'content' in chunked.get('choices')[0].get('delta'))):
+
+                    content = chunked.get('choices')[0].get('delta').get('content')
+                    if content:
+                        self._llmresponse += content
+                self._response_id = chunked.get('id')
+                self._response_model = chunked.get('model')
+                self._finish_reason = chunked.get('choices')[0].get('finish_reason')
+                self._system_fingerprint = chunked.get('system_fingerprint')
+                return chunk
+            except StopAsyncIteration:
+                # Handling exception ensure observability without disrupting operation
+                try:
+                    self._end_time = time.time()
+                    if len(self._timestamps) > 1:
+                        self._tbt = calculate_tbt(self._timestamps)
+
+                    # Format 'messages' into a single string
+                    message_prompt = self._kwargs.get("messages", "")
+                    formatted_messages = []
+                    for message in message_prompt:
+                        role = message["role"]
+                        content = message["content"]
+
+                        if isinstance(content, list):
+                            content_str_list = []
+                            for item in content:
+                                if item["type"] == "text":
+                                    content_str_list.append(f'text: {item["text"]}')
+                                elif (item["type"] == "image_url" and
+                                      not item["image_url"]["url"].startswith("data:")):
+                                    content_str_list.append(f'image_url: {item["image_url"]["url"]}')
+                            content_str = ", ".join(content_str_list)
+                            formatted_messages.append(f"{role}: {content_str}")
+                        else:
+                            formatted_messages.append(f"{role}: {content}")
+                    prompt = "\n".join(formatted_messages)
+
+                    request_model = self._kwargs.get("model", "gpt-4o")
+
+                    # Calculate tokens using input prompt and aggregated response
+                    input_tokens = general_tokens(prompt)
+                    output_tokens = general_tokens(self._llmresponse)
+
+                    # Calculate cost of the operation
+                    cost = get_chat_model_cost(request_model,
+                                                pricing_info, input_tokens,
+                                                output_tokens)
+
+                    # Set Span attributes (OTel Semconv)
+                    self._span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
+                                        SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
+                                        SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
+                                        request_model)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_SEED,
+                                        self._kwargs.get("seed", ""))
+                    self._span.set_attribute(SemanticConvetion.SERVER_PORT,
+                                        self._server_port)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_FREQUENCY_PENALTY,
+                                        self._kwargs.get("frequency_penalty", 0.0))
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MAX_TOKENS,
+                                        self._kwargs.get("max_tokens", -1))
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_PRESENCE_PENALTY,
+                                        self._kwargs.get("presence_penalty", 0.0))
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_STOP_SEQUENCES,
+                                        self._kwargs.get("stop", []))
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_TEMPERATURE,
+                                        self._kwargs.get("temperature", 1.0))
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_TOP_P,
+                                        self._kwargs.get("top_p", 1.0))
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
+                                        [self._finish_reason])
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_ID,
+                                        self._response_id)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_MODEL,
+                                        self._response_model)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
+                                        input_tokens)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_USAGE_OUTPUT_TOKENS,
+                                        output_tokens)
+                    self._span.set_attribute(SemanticConvetion.SERVER_ADDRESS,
+                                        self._server_address)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_OPENAI_RESPONSE_SYSTEM_FINGERPRINT,
+                                        self._system_fingerprint)
+                    if isinstance(self._llmresponse, str):
+                        self._span.set_attribute(SemanticConvetion.GEN_AI_OUTPUT_TYPE,
+                                        "text")
+                    else:
+                        self._span.set_attribute(SemanticConvetion.GEN_AI_OUTPUT_TYPE,
+                                        "json")
+
+                    # Set Span attributes (Extra)
+                    self._span.set_attribute(DEPLOYMENT_ENVIRONMENT,
+                                        environment)
+                    self._span.set_attribute(SERVICE_NAME,
+                                        application_name)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
+                                        True)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
+                                        input_tokens + output_tokens)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
+                                        cost)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_SERVER_TBT,
+                                        self._tbt)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_SERVER_TTFT,
+                                        self._ttft)
+                    self._span.set_attribute(SemanticConvetion.GEN_AI_SDK_VERSION,
+                                        version)
+                    if trace_content:
+                        self._span.add_event(
+                            name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
+                            attributes={
+                                SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
+                            },
+                        )
+                        self._span.add_event(
+                            name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
+                            attributes={
+                                SemanticConvetion.GEN_AI_CONTENT_COMPLETION: self._llmresponse,
+                            },
+                        )
+                    self._span.set_status(Status(StatusCode.OK))
+
+                    if disable_metrics is False:
+                        attributes = create_metrics_attributes(
+                            service_name=application_name,
+                            deployment_environment=environment,
+                            operation=SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT,
+                            system=SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE,
+                            request_model=request_model,
+                            server_address=self._server_address,
+                            server_port=self._server_port,
+                            response_model=self._response_model,
+                        )
+
+                        metrics["genai_client_usage_tokens"].record(
+                            input_tokens + output_tokens, attributes
+                        )
+                        metrics["genai_client_operation_duration"].record(
+                            self._end_time - self._start_time, attributes
+                        )
+                        metrics["genai_server_tbt"].record(
+                            self._tbt, attributes
+                        )
+                        metrics["genai_server_ttft"].record(
+                            self._ttft, attributes
+                        )
+                        metrics["genai_requests"].add(1, attributes)
+                        metrics["genai_completion_tokens"].add(output_tokens, attributes)
+                        metrics["genai_prompt_tokens"].add(input_tokens, attributes)
+                        metrics["genai_cost"].record(cost, attributes)
+
+                except Exception as e:
+                    handle_exception(self._span, e)
+                    logger.error("Error in trace creation: %s", e)
+                finally:
+                    self._span.end()
+                raise
+
     async def wrapper(wrapped, instance, args, kwargs):
         """
-        Wraps the 'chat' API call to add telemetry.
-        
+        Wraps the 'chat.completions' API call to add telemetry.
+
         This collects metrics such as execution time, cost, and token usage, and handles errors
         gracefully, adding details to the trace for observability.
 
         Args:
-            wrapped: The original 'chat' method to be wrapped.
+            wrapped: The original 'chat.completions' method to be wrapped.
             instance: The instance of the class where the original method is defined.
-            args: Positional arguments for the 'chat' method.
-            kwargs: Keyword arguments for the 'chat' method.
+            args: Positional arguments for the 'chat.completions' method.
+            kwargs: Keyword arguments for the 'chat.completions' method.
 
         Returns:
-            The response from the original 'chat' method.
+            The response from the original 'chat.completions' method.
         """
-        # pylint: disable=no-else-return
-        if kwargs.get("stream", False) is True:
-            # Special handling for streaming response to accommodate the nature of data flow
-            async def stream_generator():
-                with tracer.start_as_current_span(gen_ai_endpoint, kind= SpanKind.CLIENT) as span:
-                    # Placeholder for aggregating streaming response
-                    llmresponse = ""
-
-                    # Loop through streaming events capturing relevant details
-                    async for chunk in await wrapped(*args, **kwargs):
-                        if chunk.choices:
-                            # Collect message IDs and aggregated response from events
-                            content = chunk.choices[0].delta.content
-                            if content:
-                                llmresponse += content
-
-                        yield chunk
-                        response_id = chunk.id
-
-                    # Handling exception ensure observability without disrupting operation
-                    try:
-                        # Format 'messages' into a single string
-                        message_prompt = kwargs.get("messages", "")
-                        formatted_messages = []
-                        for message in message_prompt:
-                            role = message["role"]
-                            content = message["content"]
-
-                            if isinstance(content, list):
-                                content_str = ", ".join(
-                                    # pylint: disable=line-too-long
-                                    f'{item["type"]}: {item["text"] if "text" in item else item["image_url"]}'
-                                    if "type" in item else f'text: {item["text"]}'
-                                    for item in content
-                                )
-                                formatted_messages.append(f"{role}: {content_str}")
-                            else:
-                                formatted_messages.append(f"{role}: {content}")
-                        prompt = "\n".join(formatted_messages)
-
-                        model = kwargs.get("model", "phi3-mini-4k")
-
-                        # Calculate tokens using input prompt and aggregated response
-                        input_tokens = general_tokens(prompt)
-                        output_tokens = general_tokens(llmresponse)
-
-                        total_tokens = input_tokens + output_tokens
-                        # Calculate cost of the operation
-                        cost = get_chat_model_cost(model,
-                                                    pricing_info, input_tokens,
-                                                    output_tokens)
-
-                        # Set base span attribues
-                        span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
-                        span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
-                                            SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE)
-                        span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
-                                            SemanticConvetion.GEN_AI_TYPE_CHAT)
-                        span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
-                                            gen_ai_endpoint)
-                        span.set_attribute(SemanticConvetion.GEN_AI_ENVIRONMENT,
-                                            environment)
-                        span.set_attribute(SemanticConvetion.GEN_AI_APPLICATION_NAME,
-                                            application_name)
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
-                                            model)
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
-                                            True)
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_USER,
-                                                kwargs.get("user", ""))
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_TOP_P,
-                                            kwargs.get("top_p", 1.0))
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MAX_TOKENS,
-                                            kwargs.get("max_tokens", -1))
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_TEMPERATURE,
-                                            kwargs.get("temperature", 1.0))
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_PRESENCE_PENALTY,
-                                            kwargs.get("presence_penalty", 0.0))
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_FREQUENCY_PENALTY,
-                                            kwargs.get("frequency_penalty", 0.0))
-                        span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_SEED,
-                                            kwargs.get("seed", ""))
-                        span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_ID,
-                                            response_id)
-                        span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
-                                        input_tokens)
-                        span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
-                                            output_tokens)
-                        span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
-                                            total_tokens)
-                        span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
-                                            cost)
 
+        # Check if streaming is enabled for the API call
+        streaming = kwargs.get("stream", False)
+        server_address, server_port = set_server_address_and_port(instance, "models.github.ai", 443)
+        request_model = kwargs.get("model", "gpt-4o")
 
-                        if trace_content:
-                            span.add_event(
-                                name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
-                                attributes={
-                                    SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
-                                },
-                            )
-                            span.add_event(
-                                name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
-                                attributes={
-                                    SemanticConvetion.GEN_AI_CONTENT_COMPLETION: llmresponse,
-                                },
-                            )
+        span_name = f"{SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
 
-                        span.set_status(Status(StatusCode.OK))
-
-                        if disable_metrics is False:
-                            attributes = {
-                                TELEMETRY_SDK_NAME:
-                                    "openlit",
-                                SemanticConvetion.GEN_AI_APPLICATION_NAME:
-                                    application_name,
-                                SemanticConvetion.GEN_AI_SYSTEM:
-                                    SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE,
-                                SemanticConvetion.GEN_AI_ENVIRONMENT:
-                                    environment,
-                                SemanticConvetion.GEN_AI_TYPE:
-                                    SemanticConvetion.GEN_AI_TYPE_CHAT,
-                                SemanticConvetion.GEN_AI_REQUEST_MODEL:
-                                    model
-                            }
-
-                            metrics["genai_requests"].add(1, attributes)
-                            metrics["genai_total_tokens"].add(
-                                total_tokens, attributes
-                            )
-                            metrics["genai_completion_tokens"].add(output_tokens, attributes)
-                            metrics["genai_prompt_tokens"].add(input_tokens, attributes)
-                            metrics["genai_cost"].record(cost, attributes)
+        # pylint: disable=no-else-return
+        if streaming:
+            # Special handling for streaming response to accommodate the nature of data flow
+            awaited_wrapped = await wrapped(*args, **kwargs)
+            span = tracer.start_span(span_name, kind=SpanKind.CLIENT)
 
-                    except Exception as e:
-                        handle_exception(span, e)
-                        logger.error("Error in trace creation: %s", e)
+            return TracedAsyncStream(awaited_wrapped, span, kwargs, server_address, server_port)
 
-            return stream_generator()
+        # Handling for non-streaming responses
         else:
-            # pylint: disable=line-too-long
-            with tracer.start_as_current_span(gen_ai_endpoint, kind= SpanKind.CLIENT) as span:
+            with tracer.start_as_current_span(span_name, kind= SpanKind.CLIENT) as span:
+                start_time = time.time()
                 response = await wrapped(*args, **kwargs)
+                end_time = time.time()
+
+                response_dict = response_as_dict(response)
 
                 try:
                     # Format 'messages' into a single string
@@ -205,7 +319,6 @@ def async_complete(gen_ai_endpoint, version, environment, application_name,
 
                         if isinstance(content, list):
                             content_str = ", ".join(
-                                # pylint: disable=line-too-long
                                 f'{item["type"]}: {item["text"] if "text" in item else item["image_url"]}'
                                 if "type" in item else f'text: {item["text"]}'
                                 for item in content
@@ -215,43 +328,66 @@ def async_complete(gen_ai_endpoint, version, environment, application_name,
                             formatted_messages.append(f"{role}: {content}")
                     prompt = "\n".join(formatted_messages)
 
-                    model = kwargs.get("model", "phi3-mini-4k")
+                    input_tokens = response_dict.get('usage').get('prompt_tokens')
+                    output_tokens = response_dict.get('usage').get('completion_tokens')
 
-                    # Set base span attribues
+                    # Calculate cost of the operation
+                    cost = get_chat_model_cost(request_model,
+                                                pricing_info, input_tokens,
+                                                output_tokens)
+
+                    # Set base span attribues (OTel Semconv)
                     span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
+                    span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
+                                        SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT)
                     span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
                                         SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE)
-                    span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
-                                        SemanticConvetion.GEN_AI_TYPE_CHAT)
-                    span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
-                                        gen_ai_endpoint)
-                    span.set_attribute(SemanticConvetion.GEN_AI_ENVIRONMENT,
-                                        environment)
-                    span.set_attribute(SemanticConvetion.GEN_AI_APPLICATION_NAME,
-                                        application_name)
                     span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
-                                        model)
-                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
-                                        False)
-                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_USER,
-                                            kwargs.get("user", ""))
-                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_TOP_P,
-                                        kwargs.get("top_p", 1.0))
+                                        request_model)
+                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_SEED,
+                                        kwargs.get("seed", ""))
+                    span.set_attribute(SemanticConvetion.SERVER_PORT,
+                                        server_port)
+                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_FREQUENCY_PENALTY,
+                                        kwargs.get("frequency_penalty", 0.0))
                     span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MAX_TOKENS,
                                         kwargs.get("max_tokens", -1))
-                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_TEMPERATURE,
-                                        kwargs.get("temperature", 1.0))
                     span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_PRESENCE_PENALTY,
                                         kwargs.get("presence_penalty", 0.0))
-                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_FREQUENCY_PENALTY,
-                                        kwargs.get("frequency_penalty", 0.0))
-                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_SEED,
-                                        kwargs.get("seed", ""))
+                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_STOP_SEQUENCES,
+                                        kwargs.get("stop", []))
+                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_TEMPERATURE,
+                                        kwargs.get("temperature", 1.0))
+                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_TOP_P,
+                                        kwargs.get("top_p", 1.0))
                     span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_ID,
-                                        response.id)
-                    span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
-                                        [response.choices[0]["finish_reason"]])
+                                        response_dict.get("id"))
+                    span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_MODEL,
+                                        response_dict.get('model'))
+                    span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
+                                        input_tokens)
+                    span.set_attribute(SemanticConvetion.GEN_AI_USAGE_OUTPUT_TOKENS,
+                                        output_tokens)
+                    span.set_attribute(SemanticConvetion.SERVER_ADDRESS,
+                                        server_address)
+                    span.set_attribute(SemanticConvetion.GEN_AI_OPENAI_RESPONSE_SYSTEM_FINGERPRINT,
+                                        response_dict.get('system_fingerprint'))
 
+                    # Set base span attribues (Extras)
+                    span.set_attribute(DEPLOYMENT_ENVIRONMENT,
+                                        environment)
+                    span.set_attribute(SERVICE_NAME,
+                                        application_name)
+                    span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_IS_STREAM,
+                                        False)
+                    span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
+                                        input_tokens + output_tokens)
+                    span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
+                                        cost)
+                    span.set_attribute(SemanticConvetion.GEN_AI_SERVER_TTFT,
+                                        end_time - start_time)
+                    span.set_attribute(SemanticConvetion.GEN_AI_SDK_VERSION,
+                                        version)
                     if trace_content:
                         span.add_event(
                             name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
@@ -259,49 +395,53 @@ def async_complete(gen_ai_endpoint, version, environment, application_name,
                                 SemanticConvetion.GEN_AI_CONTENT_PROMPT: prompt,
                             },
                         )
-                        span.add_event(
-                            name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
-                            attributes={
-                                SemanticConvetion.GEN_AI_CONTENT_COMPLETION: response.choices[0].message.content,
-                            },
-                        )
 
-                    input_tokens = response.usage.prompt_tokens
-                    output_tokens = response.usage.completion_tokens
-                    total_tokens = response.usage.total_tokens
-                    # Calculate cost of the operation
-                    cost = get_chat_model_cost(model,
-                                                pricing_info, input_tokens, output_tokens)
+                    for i in range(kwargs.get('n',1)):
+                        span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_FINISH_REASON,
+                                           [response_dict.get('choices')[i].get('finish_reason')])
+                        if trace_content:
+                            span.add_event(
+                                name=SemanticConvetion.GEN_AI_CONTENT_COMPLETION_EVENT,
+                                attributes={
+                                    # pylint: disable=line-too-long
+                                    SemanticConvetion.GEN_AI_CONTENT_COMPLETION: str(response_dict.get('choices')[i].get('message').get('content')),
+                                },
+                            )
+                        if kwargs.get('tools'):
+                            span.set_attribute(SemanticConvetion.GEN_AI_TOOL_CALLS,
+                                            str(response_dict.get('choices')[i].get('message').get('tool_calls')))
 
-                    span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
-                                        input_tokens)
-                    span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COMPLETION_TOKENS,
-                                        output_tokens)
-                    span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
-                                        total_tokens)
-                    span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
-                                        cost)
+                        if isinstance(response_dict.get('choices')[i].get('message').get('content'), str):
+                            span.set_attribute(SemanticConvetion.GEN_AI_OUTPUT_TYPE,
+                                            "text")
+                        elif response_dict.get('choices')[i].get('message').get('content') is not None:
+                            span.set_attribute(SemanticConvetion.GEN_AI_OUTPUT_TYPE,
+                                            "json")
 
                     span.set_status(Status(StatusCode.OK))
 
                     if disable_metrics is False:
-                        attributes = {
-                            TELEMETRY_SDK_NAME:
-                                "openlit",
-                            SemanticConvetion.GEN_AI_APPLICATION_NAME:
-                                application_name,
-                            SemanticConvetion.GEN_AI_SYSTEM:
-                                SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE,
-                            SemanticConvetion.GEN_AI_ENVIRONMENT:
-                                environment,
-                            SemanticConvetion.GEN_AI_TYPE:
-                                SemanticConvetion.GEN_AI_TYPE_CHAT,
-                            SemanticConvetion.GEN_AI_REQUEST_MODEL:
-                                model
-                        }
+                        attributes = create_metrics_attributes(
+                            service_name=application_name,
+                            deployment_environment=environment,
+                            operation=SemanticConvetion.GEN_AI_OPERATION_TYPE_CHAT,
+                            system=SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE,
+                            request_model=request_model,
+                            server_address=server_address,
+                            server_port=server_port,
+                            response_model=response_dict.get('model'),
+                        )
 
+                        metrics["genai_client_usage_tokens"].record(
+                            input_tokens + output_tokens, attributes
+                        )
+                        metrics["genai_client_operation_duration"].record(
+                            end_time - start_time, attributes
+                        )
+                        metrics["genai_server_ttft"].record(
+                            end_time - start_time, attributes
+                        )
                         metrics["genai_requests"].add(1, attributes)
-                        metrics["genai_total_tokens"].add(total_tokens, attributes)
                         metrics["genai_completion_tokens"].add(output_tokens, attributes)
                         metrics["genai_prompt_tokens"].add(input_tokens, attributes)
                         metrics["genai_cost"].record(cost, attributes)
@@ -318,18 +458,17 @@ def async_complete(gen_ai_endpoint, version, environment, application_name,
 
     return wrapper
 
-def async_embedding(gen_ai_endpoint, version, environment, application_name,
+def async_embedding(version, environment, application_name,
               tracer, pricing_info, trace_content, metrics, disable_metrics):
     """
     Generates a telemetry wrapper for embeddings to collect metrics.
     
     Args:
-        gen_ai_endpoint: Endpoint identifier for logging and tracing.
         version: Version of the monitoring package.
         environment: Deployment environment (e.g., production, staging).
-        application_name: Name of the application using the Azure AI Inference API.
+        application_name: Name of the application using the Azure Inference API.
         tracer: OpenTelemetry tracer for creating spans.
-        pricing_info: Information used for calculating the cost of Azure AI Inference usage.
+        pricing_info: Information used for calculating the cost of Azure Inference usage.
         trace_content: Flag indicating whether to trace the actual content.
     
     Returns:
@@ -353,69 +492,84 @@ def async_embedding(gen_ai_endpoint, version, environment, application_name,
             The response from the original 'embeddings' method.
         """
 
-        with tracer.start_as_current_span(gen_ai_endpoint, kind= SpanKind.CLIENT) as span:
+        server_address, server_port = set_server_address_and_port(instance, "models.github.ai", 443)
+        request_model = kwargs.get("model", "text-embedding-ada-002")
+
+        span_name = f"{SemanticConvetion.GEN_AI_OPERATION_TYPE_EMBEDDING} {request_model}"
+
+        with tracer.start_as_current_span(span_name, kind= SpanKind.CLIENT) as span:
+            start_time = time.time()
             response = await wrapped(*args, **kwargs)
+            end_time = time.time()
 
+            response_dict = response_as_dict(response)
             try:
+                input_tokens = response_dict.get('usage').get('prompt_tokens')
+
                 # Calculate cost of the operation
-                cost = get_embed_model_cost(kwargs.get("model", "text-embedding-ada-002"),
-                                            pricing_info, response.usage.prompt_tokens)
+                cost = get_embed_model_cost(request_model,
+                                    pricing_info, input_tokens)
 
-                # Set Span attributes
+                # Set Span attributes (OTel Semconv)
                 span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
+                span.set_attribute(SemanticConvetion.GEN_AI_OPERATION,
+                                    SemanticConvetion.GEN_AI_OPERATION_TYPE_EMBEDDING)
                 span.set_attribute(SemanticConvetion.GEN_AI_SYSTEM,
                                     SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE)
-                span.set_attribute(SemanticConvetion.GEN_AI_TYPE,
-                                    SemanticConvetion.GEN_AI_TYPE_EMBEDDING)
-                span.set_attribute(SemanticConvetion.GEN_AI_ENDPOINT,
-                                    gen_ai_endpoint)
-                span.set_attribute(SemanticConvetion.GEN_AI_ENVIRONMENT,
+                span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
+                                    request_model)
+                span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_ENCODING_FORMATS,
+                                    [kwargs.get('encoding_format', 'float')])
+                span.set_attribute(SemanticConvetion.GEN_AI_RESPONSE_MODEL,
+                                    request_model)
+                span.set_attribute(SemanticConvetion.SERVER_ADDRESS,
+                                    server_address)
+                span.set_attribute(SemanticConvetion.SERVER_PORT,
+                                    server_port)
+                span.set_attribute(SemanticConvetion.GEN_AI_USAGE_INPUT_TOKENS,
+                                    input_tokens)
+
+                # Set Span attributes (Extras)
+                span.set_attribute(DEPLOYMENT_ENVIRONMENT,
                                     environment)
-                span.set_attribute(SemanticConvetion.GEN_AI_APPLICATION_NAME,
+                span.set_attribute(SERVICE_NAME,
                                     application_name)
-                span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_MODEL,
-                                    kwargs.get("model", "text-embedding-ada-002"))
-                span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_EMBEDDING_FORMAT,
-                                    kwargs.get("encoding_format", "float"))
-                span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_EMBEDDING_DIMENSION,
-                                    kwargs.get("dimensions", ""))
-                span.set_attribute(SemanticConvetion.GEN_AI_REQUEST_USER,
-                                    kwargs.get("user", ""))
-                span.set_attribute(SemanticConvetion.GEN_AI_USAGE_PROMPT_TOKENS,
-                                    response.usage.prompt_tokens)
                 span.set_attribute(SemanticConvetion.GEN_AI_USAGE_TOTAL_TOKENS,
-                                    response.usage.total_tokens)
+                                    input_tokens)
                 span.set_attribute(SemanticConvetion.GEN_AI_USAGE_COST,
                                     cost)
+                span.set_attribute(SemanticConvetion.GEN_AI_SDK_VERSION,
+                                    version)
+
                 if trace_content:
                     span.add_event(
                         name=SemanticConvetion.GEN_AI_CONTENT_PROMPT_EVENT,
                         attributes={
-                            SemanticConvetion.GEN_AI_CONTENT_PROMPT: kwargs.get("input", ""),
+                            SemanticConvetion.GEN_AI_CONTENT_PROMPT: str(kwargs.get("input", "")),
                         },
                     )
 
                 span.set_status(Status(StatusCode.OK))
 
                 if disable_metrics is False:
-                    attributes = {
-                        TELEMETRY_SDK_NAME:
-                            "openlit",
-                        SemanticConvetion.GEN_AI_APPLICATION_NAME:
-                            application_name,
-                        SemanticConvetion.GEN_AI_SYSTEM:
-                            SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE,
-                        SemanticConvetion.GEN_AI_ENVIRONMENT:
-                            environment,
-                        SemanticConvetion.GEN_AI_TYPE:
-                            SemanticConvetion.GEN_AI_TYPE_EMBEDDING,
-                        SemanticConvetion.GEN_AI_REQUEST_MODEL:
-                            kwargs.get("model", "text-embedding-ada-002")
-                    }
-
+                    attributes = create_metrics_attributes(
+                        service_name=application_name,
+                        deployment_environment=environment,
+                        operation=SemanticConvetion.GEN_AI_OPERATION_TYPE_EMBEDDING,
+                        system=SemanticConvetion.GEN_AI_SYSTEM_AZURE_AI_INFERENCE,
+                        request_model=request_model,
+                        server_address=server_address,
+                        server_port=server_port,
+                        response_model=request_model,
+                    )
+                    metrics["genai_client_usage_tokens"].record(
+                            input_tokens, attributes
+                        )
+                    metrics["genai_client_operation_duration"].record(
+                        end_time - start_time, attributes
+                    )
                     metrics["genai_requests"].add(1, attributes)
-                    metrics["genai_total_tokens"].add(response.usage.total_tokens, attributes)
-                    metrics["genai_prompt_tokens"].add(response.usage.prompt_tokens, attributes)
+                    metrics["genai_prompt_tokens"].add(input_tokens, attributes)
                     metrics["genai_cost"].record(cost, attributes)
 
                 # Return original response