Skip to content

Same agent Smaller bill. The tools and infrastructure to get more work from your AI. Fewer tokens. Less compute. Lower cost.

[Feat]: Refactor Groq OpenTelemetry Monitoring Instrumentation (#774)

Patcher1c7987f6ee0a
diff --git a/sdk/python/pyproject.toml b/sdk/python/pyproject.toml
index 7d76844a..4e0ec050 100644
--- a/sdk/python/pyproject.toml
+++ b/sdk/python/pyproject.toml
@@ -1,6 +1,6 @@
[tool.poetry]
name = "openlit"
-version = "1.34.10"
+version = "1.34.11"
description = "OpenTelemetry-native Auto instrumentation library for monitoring LLM Applications and GPUs, facilitating the integration of observability into your GenAI-driven projects"
authors = ["OpenLIT"]
license = "Apache-2.0"
diff --git a/sdk/python/src/openlit/__helpers.py b/sdk/python/src/openlit/__helpers.py
index bb47eb5a..3b592fe1 100644
--- a/sdk/python/src/openlit/__helpers.py
+++ b/sdk/python/src/openlit/__helpers.py
@@ -379,3 +379,26 @@ def record_completion_metrics(metrics, gen_ai_operation, gen_ai_system, server_a
metrics["genai_completion_tokens"].add(output_tokens, attributes)
metrics["genai_prompt_tokens"].add(input_tokens, attributes)
metrics["genai_cost"].record(cost, attributes)
+
+def record_embedding_metrics(metrics, gen_ai_operation, gen_ai_system, server_address, server_port,
+ request_model, response_model, environment, application_name, start_time, end_time,
+ input_tokens, cost):
+ """
+ Record embedding-specific metrics for the operation.
+ """
+
+ attributes = create_metrics_attributes(
+ operation=gen_ai_operation,
+ system=gen_ai_system,
+ server_address=server_address,
+ server_port=server_port,
+ request_model=request_model,
+ response_model=response_model,
+ service_name=application_name,
+ deployment_environment=environment,
+ )
+ 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_prompt_tokens"].add(input_tokens, attributes)
+ metrics["genai_cost"].record(cost, attributes)
diff --git a/sdk/python/src/openlit/instrumentation/groq/__init__.py b/sdk/python/src/openlit/instrumentation/groq/__init__.py
index 1605bfed..25578aed 100644
--- a/sdk/python/src/openlit/instrumentation/groq/__init__.py
+++ b/sdk/python/src/openlit/instrumentation/groq/__init__.py
@@ -1,4 +1,3 @@
-# pylint: disable=useless-return, bad-staticmethod-argument, disable=duplicate-code
"""Initializer of Auto Instrumentation of Groq Functions"""
from typing import Collection
@@ -13,15 +12,15 @@ _instruments = ("groq >= 0.5.0",)
class GroqInstrumentor(BaseInstrumentor):
"""
- An instrumentor for Groq's client library.
+ An instrumentor for Groq client library.
"""
def instrumentation_dependencies(self) -> Collection[str]:
return _instruments
def _instrument(self, **kwargs):
- application_name = kwargs.get("application_name", "default_application")
- environment = kwargs.get("environment", "default_environment")
+ application_name = kwargs.get("application_name", "default")
+ environment = kwargs.get("environment", "default")
tracer = kwargs.get("tracer")
metrics = kwargs.get("metrics_dict")
pricing_info = kwargs.get("pricing_info", {})
@@ -29,22 +28,21 @@ class GroqInstrumentor(BaseInstrumentor):
disable_metrics = kwargs.get("disable_metrics")
version = importlib.metadata.version("groq")
- #sync
+ # Chat completions
wrap_function_wrapper(
"groq.resources.chat.completions",
"Completions.create",
chat(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics),
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),
)
- #async
+ # Chat completions
wrap_function_wrapper(
"groq.resources.chat.completions",
"AsyncCompletions.create",
async_chat(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics),
+ tracer, pricing_info, capture_message_content, metrics, disable_metrics),
)
def _uninstrument(self, **kwargs):
- # Proper uninstrumentation logic to revert patched methods
pass
diff --git a/sdk/python/src/openlit/instrumentation/groq/async_groq.py b/sdk/python/src/openlit/instrumentation/groq/async_groq.py
index 397dcebd..b66b5fb6 100644
--- a/sdk/python/src/openlit/instrumentation/groq/async_groq.py
+++ b/sdk/python/src/openlit/instrumentation/groq/async_groq.py
@@ -1,57 +1,36 @@
"""
-Module for monitoring Groq API calls.
+Module for monitoring Groq API calls (async version).
"""
-import logging
import time
-from opentelemetry.trace import SpanKind, Status, StatusCode
-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT
+from opentelemetry.trace import SpanKind
from openlit.__helpers import (
- get_chat_model_cost,
handle_exception,
- response_as_dict,
- calculate_ttft,
- calculate_tbt,
- create_metrics_attributes,
set_server_address_and_port
)
+from openlit.instrumentation.groq.utils import (
+ process_chunk,
+ process_streaming_chat_response,
+ process_chat_response
+)
from openlit.semcov import SemanticConvention
-# Initialize logger for logging potential issues and operations
-logger = logging.getLogger(__name__)
-
-def async_chat(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics):
+def async_chat(version, environment, application_name, tracer, pricing_info,
+ capture_message_content, metrics, disable_metrics):
"""
- Generates a telemetry wrapper for chat completions to collect metrics.
-
- Args:
- version: Version of the monitoring package.
- environment: Deployment environment (e.g., production, staging).
- application_name: Name of the application using the Groq API.
- tracer: OpenTelemetry tracer for creating spans.
- pricing_info: Information used for calculating the cost of Groq usage.
- capture_message_content: Flag indicating whether to trace the actual content.
-
- Returns:
- A function that wraps the chat completions method to add telemetry.
+ Generates a telemetry wrapper for GenAI function call
"""
class TracedAsyncStream:
"""
- Wrapper for streaming responses to collect metrics and trace data.
- Wraps the 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.
+ Wrapper for async streaming responses to collect telemetry.
"""
+
def __init__(
self,
wrapped,
span,
+ span_name,
kwargs,
server_address,
server_port,
@@ -59,15 +38,15 @@ def async_chat(version, environment, application_name,
):
self.__wrapped__ = wrapped
self._span = span
- # Placeholder for aggregating streaming response
+ self._span_name = span_name
self._llmresponse = ""
self._response_id = ""
self._response_model = ""
self._finish_reason = ""
+ self._tools = None
self._system_fingerprint = ""
self._input_tokens = 0
self._output_tokens = 0
-
self._args = args
self._kwargs = kwargs
self._start_time = time.time()
@@ -95,373 +74,70 @@ def async_chat(version, environment, application_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
-
- if chunked.get('usage'):
- self._input_tokens = chunked.get('usage').get('prompt_tokens')
- self._output_tokens = chunked.get('usage').get('completion_tokens')
- 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')
+ process_chunk(self, chunk)
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 cost of the operation
- cost = get_chat_model_cost(request_model,
- pricing_info, self._input_tokens,
- self._output_tokens)
-
- # Set Span attributes (OTel Semconv)
- self._span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
- self._span.set_attribute(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT)
- self._span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_GROQ)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED,
- self._kwargs.get("seed", ""))
- self._span.set_attribute(SemanticConvention.SERVER_PORT,
- self._server_port)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY,
- self._kwargs.get("frequency_penalty", 0.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS,
- self._kwargs.get("max_completion_tokens", -1))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY,
- self._kwargs.get("presence_penalty", 0.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES,
- self._kwargs.get("stop", []))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE,
- self._kwargs.get("temperature", 1.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P,
- self._kwargs.get("top_p", 1.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON,
- [self._finish_reason])
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID,
- self._response_id)
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- self._response_model)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- self._input_tokens)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS,
- self._output_tokens)
- self._span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- self._server_address)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SERVICE_TIER,
- self._kwargs.get("service_tier", "on_demand"))
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_SYSTEM_FINGERPRINT,
- self._system_fingerprint)
- if isinstance(self._llmresponse, str):
- self._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- "text")
- else:
- self._span.set_attribute(SemanticConvention.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(SemanticConvention.GEN_AI_REQUEST_USER,
- self._kwargs.get("user", ""))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM,
- True)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- self._input_tokens + self._output_tokens)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- self._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TBT,
- self._tbt)
- self._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT,
- self._ttft)
- self._span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
- if capture_message_content:
- self._span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
- },
- )
- self._span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
- attributes={
- SemanticConvention.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=SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT,
- system=SemanticConvention.GEN_AI_SYSTEM_GROQ,
- 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(
- self._input_tokens + self._output_tokens, attributes
- )
- metrics["genai_client_operation_duration"].record(
- self._end_time - self._start_time, attributes
- )
- metrics["genai_server_tbt"].record(
- self._tbt, attributes
+ with tracer.start_as_current_span(self._span_name, kind= SpanKind.CLIENT) as self._span:
+ process_streaming_chat_response(
+ self,
+ pricing_info=pricing_info,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version
)
- metrics["genai_server_ttft"].record(
- self._ttft, attributes
- )
- metrics["genai_requests"].add(1, attributes)
- metrics["genai_completion_tokens"].add(self._output_tokens, attributes)
- metrics["genai_prompt_tokens"].add(self._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.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.completions' method to be wrapped.
- instance: The instance of the class where the original method is defined.
- args: Positional arguments for the 'chat.completions' method.
- kwargs: Keyword arguments for the 'chat.completions' method.
-
- Returns:
- The response from the original 'chat.completions' method.
+ Wraps the GenAI function call.
"""
-
# Check if streaming is enabled for the API call
streaming = kwargs.get("stream", False)
server_address, server_port = set_server_address_and_port(instance, "api.groq.com", 443)
- request_model = kwargs.get("model", "gpt-4o")
+ request_model = kwargs.get("model", "mixtral-8x7b-32768")
span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
- # pylint: disable=no-else-return
if streaming:
- # Special handling for streaming response to accommodate the nature of data flow
+ # Special handling for streaming response
awaited_wrapped = await wrapped(*args, **kwargs)
span = tracer.start_span(span_name, kind=SpanKind.CLIENT)
-
- return TracedAsyncStream(awaited_wrapped, span, kwargs, server_address, server_port)
-
- # Handling for non-streaming responses
+ return TracedAsyncStream(awaited_wrapped, span, span_name, kwargs, server_address, server_port)
else:
- with tracer.start_as_current_span(span_name, kind= SpanKind.CLIENT) as span:
+ # Handling for non-streaming responses
+ 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
- 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(
- 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)
-
- input_tokens = response_dict.get('usage').get('prompt_tokens')
- output_tokens = response_dict.get('usage').get('completion_tokens')
-
- # 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(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT)
- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_GROQ)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED,
- kwargs.get("seed", ""))
- span.set_attribute(SemanticConvention.SERVER_PORT,
- server_port)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY,
- kwargs.get("frequency_penalty", 0.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS,
- kwargs.get("max_completion_tokens", -1))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY,
- kwargs.get("presence_penalty", 0.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES,
- kwargs.get("stop", []))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE,
- kwargs.get("temperature", 1.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P,
- kwargs.get("top_p", 1.0))
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID,
- response_dict.get("id"))
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- response_dict.get('model'))
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- input_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS,
- output_tokens)
- span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- server_address)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SERVICE_TIER,
- kwargs.get("service_tier", "on_demand"))
- span.set_attribute(SemanticConvention.GEN_AI_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(SemanticConvention.GEN_AI_REQUEST_USER,
- kwargs.get("user", ""))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM,
- False)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- input_tokens + output_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT,
- end_time - start_time)
- span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
- },
- )
-
- for i in range(kwargs.get('n',1)):
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON,
- [response_dict.get('choices')[i].get('finish_reason')])
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
- attributes={
- # pylint: disable=line-too-long
- SemanticConvention.GEN_AI_CONTENT_COMPLETION: str(response_dict.get('choices')[i].get('message').get('content')),
- },
- )
- if kwargs.get('tools'):
- span.set_attribute(SemanticConvention.GEN_AI_TOOL_CALLS,
- str(response_dict.get('choices')[i].get('message').get('tool_calls')))
-
- if isinstance(response_dict.get('choices')[i].get('message').get('content'), str):
- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- "text")
- elif response_dict.get('choices')[i].get('message').get('content') is not None:
- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- "json")
-
- span.set_status(Status(StatusCode.OK))
-
- if disable_metrics is False:
- attributes = create_metrics_attributes(
- service_name=application_name,
- deployment_environment=environment,
- operation=SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT,
- system=SemanticConvention.GEN_AI_SYSTEM_GROQ,
- 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_completion_tokens"].add(output_tokens, attributes)
- metrics["genai_prompt_tokens"].add(input_tokens, attributes)
- metrics["genai_cost"].record(cost, attributes)
-
- # Return original response
- return response
+ response = process_chat_response(
+ response=response,
+ request_model=request_model,
+ pricing_info=pricing_info,
+ server_port=server_port,
+ server_address=server_address,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ start_time=start_time,
+ span=span,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version,
+ **kwargs
+ )
except Exception as e:
handle_exception(span, e)
- logger.error("Error in trace creation: %s", e)
- # Return original response
- return response
+ return response
return wrapper
diff --git a/sdk/python/src/openlit/instrumentation/groq/groq.py b/sdk/python/src/openlit/instrumentation/groq/groq.py
index 85d23483..15da7708 100644
--- a/sdk/python/src/openlit/instrumentation/groq/groq.py
+++ b/sdk/python/src/openlit/instrumentation/groq/groq.py
@@ -2,56 +2,35 @@
Module for monitoring Groq API calls.
"""
-import logging
import time
-from opentelemetry.trace import SpanKind, Status, StatusCode
-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT
+from opentelemetry.trace import SpanKind
from openlit.__helpers import (
- get_chat_model_cost,
handle_exception,
- response_as_dict,
- calculate_ttft,
- calculate_tbt,
- create_metrics_attributes,
set_server_address_and_port
)
+from openlit.instrumentation.groq.utils import (
+ process_chunk,
+ process_streaming_chat_response,
+ process_chat_response
+)
from openlit.semcov import SemanticConvention
-# Initialize logger for logging potential issues and operations
-logger = logging.getLogger(__name__)
-
-def chat(version, environment, application_name,
- tracer, pricing_info, capture_message_content, metrics, disable_metrics):
+def chat(version, environment, application_name, tracer, pricing_info,
+ capture_message_content, metrics, disable_metrics):
"""
- Generates a telemetry wrapper for chat completions to collect metrics.
-
- Args:
- version: Version of the monitoring package.
- environment: Deployment environment (e.g., production, staging).
- application_name: Name of the application using the Groq API.
- tracer: OpenTelemetry tracer for creating spans.
- pricing_info: Information used for calculating the cost of Groq usage.
- capture_message_content: Flag indicating whether to trace the actual content.
-
- Returns:
- A function that wraps the chat completions method to add telemetry.
+ Generates a telemetry wrapper for GenAI function call
"""
class TracedSyncStream:
"""
- Wrapper for streaming responses to collect metrics and trace data.
- Wraps the 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.
+ Wrapper for streaming responses to collect telemetry.
"""
+
def __init__(
self,
wrapped,
span,
+ span_name,
kwargs,
server_address,
server_port,
@@ -59,15 +38,15 @@ def chat(version, environment, application_name,
):
self.__wrapped__ = wrapped
self._span = span
- # Placeholder for aggregating streaming response
+ self._span_name = span_name
self._llmresponse = ""
self._response_id = ""
self._response_model = ""
self._finish_reason = ""
+ self._tools = None
self._system_fingerprint = ""
self._input_tokens = 0
self._output_tokens = 0
-
self._args = args
self._kwargs = kwargs
self._start_time = time.time()
@@ -95,373 +74,70 @@ def chat(version, environment, application_name,
def __next__(self):
try:
chunk = self.__wrapped__.__next__()
- 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
-
- if chunked.get('usage'):
- self._input_tokens = chunked.get('usage').get('prompt_tokens')
- self._output_tokens = chunked.get('usage').get('completion_tokens')
- 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')
+ process_chunk(self, chunk)
return chunk
except StopIteration:
- # 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 cost of the operation
- cost = get_chat_model_cost(request_model,
- pricing_info, self._input_tokens,
- self._output_tokens)
-
- # Set Span attributes (OTel Semconv)
- self._span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
- self._span.set_attribute(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT)
- self._span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_GROQ)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED,
- self._kwargs.get("seed", ""))
- self._span.set_attribute(SemanticConvention.SERVER_PORT,
- self._server_port)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY,
- self._kwargs.get("frequency_penalty", 0.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS,
- self._kwargs.get("max_completion_tokens", -1))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY,
- self._kwargs.get("presence_penalty", 0.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES,
- self._kwargs.get("stop", []))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE,
- self._kwargs.get("temperature", 1.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P,
- self._kwargs.get("top_p", 1.0))
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON,
- [self._finish_reason])
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID,
- self._response_id)
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- self._response_model)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- self._input_tokens)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS,
- self._output_tokens)
- self._span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- self._server_address)
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SERVICE_TIER,
- self._kwargs.get("service_tier", "on_demand"))
- self._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_SYSTEM_FINGERPRINT,
- self._system_fingerprint)
- if isinstance(self._llmresponse, str):
- self._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- "text")
- else:
- self._span.set_attribute(SemanticConvention.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(SemanticConvention.GEN_AI_REQUEST_USER,
- self._kwargs.get("user", ""))
- self._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM,
- True)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- self._input_tokens + self._output_tokens)
- self._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- self._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TBT,
- self._tbt)
- self._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT,
- self._ttft)
- self._span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
- if capture_message_content:
- self._span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
- },
- )
- self._span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
- attributes={
- SemanticConvention.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=SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT,
- system=SemanticConvention.GEN_AI_SYSTEM_GROQ,
- 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(
- self._input_tokens + self._output_tokens, attributes
- )
- metrics["genai_client_operation_duration"].record(
- self._end_time - self._start_time, attributes
- )
- metrics["genai_server_tbt"].record(
- self._tbt, attributes
+ with tracer.start_as_current_span(self._span_name, kind= SpanKind.CLIENT) as self._span:
+ process_streaming_chat_response(
+ self,
+ pricing_info=pricing_info,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version
)
- metrics["genai_server_ttft"].record(
- self._ttft, attributes
- )
- metrics["genai_requests"].add(1, attributes)
- metrics["genai_completion_tokens"].add(self._output_tokens, attributes)
- metrics["genai_prompt_tokens"].add(self._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
def wrapper(wrapped, instance, args, kwargs):
"""
- 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.completions' method to be wrapped.
- instance: The instance of the class where the original method is defined.
- args: Positional arguments for the 'chat.completions' method.
- kwargs: Keyword arguments for the 'chat.completions' method.
-
- Returns:
- The response from the original 'chat.completions' method.
+ Wraps the GenAI function call.
"""
-
# Check if streaming is enabled for the API call
streaming = kwargs.get("stream", False)
server_address, server_port = set_server_address_and_port(instance, "api.groq.com", 443)
- request_model = kwargs.get("model", "gpt-4o")
+ request_model = kwargs.get("model", "mixtral-8x7b-32768")
span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
- # pylint: disable=no-else-return
if streaming:
- # Special handling for streaming response to accommodate the nature of data flow
+ # Special handling for streaming response
awaited_wrapped = wrapped(*args, **kwargs)
span = tracer.start_span(span_name, kind=SpanKind.CLIENT)
-
- return TracedSyncStream(awaited_wrapped, span, kwargs, server_address, server_port)
-
- # Handling for non-streaming responses
+ return TracedSyncStream(awaited_wrapped, span, span_name, kwargs, server_address, server_port)
else:
- with tracer.start_as_current_span(span_name, kind= SpanKind.CLIENT) as span:
+ # Handling for non-streaming responses
+ with tracer.start_as_current_span(span_name, kind=SpanKind.CLIENT) as span:
start_time = time.time()
response = wrapped(*args, **kwargs)
- end_time = time.time()
-
- response_dict = response_as_dict(response)
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(
- 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)
-
- input_tokens = response_dict.get('usage').get('prompt_tokens')
- output_tokens = response_dict.get('usage').get('completion_tokens')
-
- # 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(SemanticConvention.GEN_AI_OPERATION,
- SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT)
- span.set_attribute(SemanticConvention.GEN_AI_SYSTEM,
- SemanticConvention.GEN_AI_SYSTEM_GROQ)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL,
- request_model)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED,
- kwargs.get("seed", ""))
- span.set_attribute(SemanticConvention.SERVER_PORT,
- server_port)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY,
- kwargs.get("frequency_penalty", 0.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS,
- kwargs.get("max_completion_tokens", -1))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY,
- kwargs.get("presence_penalty", 0.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES,
- kwargs.get("stop", []))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE,
- kwargs.get("temperature", 1.0))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P,
- kwargs.get("top_p", 1.0))
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID,
- response_dict.get("id"))
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL,
- response_dict.get('model'))
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS,
- input_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS,
- output_tokens)
- span.set_attribute(SemanticConvention.SERVER_ADDRESS,
- server_address)
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SERVICE_TIER,
- kwargs.get("service_tier", "on_demand"))
- span.set_attribute(SemanticConvention.GEN_AI_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(SemanticConvention.GEN_AI_REQUEST_USER,
- kwargs.get("user", ""))
- span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM,
- False)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_TOTAL_TOKENS,
- input_tokens + output_tokens)
- span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST,
- cost)
- span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT,
- end_time - start_time)
- span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION,
- version)
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
- attributes={
- SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
- },
- )
-
- for i in range(kwargs.get('n',1)):
- span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON,
- [response_dict.get('choices')[i].get('finish_reason')])
- if capture_message_content:
- span.add_event(
- name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
- attributes={
- # pylint: disable=line-too-long
- SemanticConvention.GEN_AI_CONTENT_COMPLETION: str(response_dict.get('choices')[i].get('message').get('content')),
- },
- )
- if kwargs.get('tools'):
- span.set_attribute(SemanticConvention.GEN_AI_TOOL_CALLS,
- str(response_dict.get('choices')[i].get('message').get('tool_calls')))
-
- if isinstance(response_dict.get('choices')[i].get('message').get('content'), str):
- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- "text")
- elif response_dict.get('choices')[i].get('message').get('content') is not None:
- span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE,
- "json")
-
- span.set_status(Status(StatusCode.OK))
-
- if disable_metrics is False:
- attributes = create_metrics_attributes(
- service_name=application_name,
- deployment_environment=environment,
- operation=SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT,
- system=SemanticConvention.GEN_AI_SYSTEM_GROQ,
- 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_completion_tokens"].add(output_tokens, attributes)
- metrics["genai_prompt_tokens"].add(input_tokens, attributes)
- metrics["genai_cost"].record(cost, attributes)
-
- # Return original response
- return response
+ response = process_chat_response(
+ response=response,
+ request_model=request_model,
+ pricing_info=pricing_info,
+ server_port=server_port,
+ server_address=server_address,
+ environment=environment,
+ application_name=application_name,
+ metrics=metrics,
+ start_time=start_time,
+ span=span,
+ capture_message_content=capture_message_content,
+ disable_metrics=disable_metrics,
+ version=version,
+ **kwargs
+ )
except Exception as e:
handle_exception(span, e)
- logger.error("Error in trace creation: %s", e)
- # Return original response
- return response
+ return response
return wrapper
diff --git a/sdk/python/src/openlit/instrumentation/groq/utils.py b/sdk/python/src/openlit/instrumentation/groq/utils.py
new file mode 100644
index 00000000..23c9980b
--- /dev/null
+++ b/sdk/python/src/openlit/instrumentation/groq/utils.py
@@ -0,0 +1,199 @@
+"""
+Groq OpenTelemetry instrumentation utility functions
+"""
+import time
+
+from opentelemetry.trace import Status, StatusCode
+
+from openlit.__helpers import (
+ calculate_ttft,
+ response_as_dict,
+ calculate_tbt,
+ get_chat_model_cost,
+ common_span_attributes,
+ record_completion_metrics,
+)
+from openlit.semcov import SemanticConvention
+
+def format_content(messages):
+ """
+ Process a list of messages to extract content.
+ """
+
+ formatted_messages = []
+ for message in messages:
+ role = message["role"]
+ content = message["content"]
+
+ if isinstance(content, list):
+ content_str = ", ".join(
+ 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}")
+
+ return "\n".join(formatted_messages)
+
+def process_chunk(scope, chunk):
+ """
+ Process a chunk of response data and update state.
+ """
+
+ end_time = time.time()
+ # Record the timestamp for the current chunk
+ scope._timestamps.append(end_time)
+
+ if len(scope._timestamps) == 1:
+ # Calculate time to first chunk
+ scope._ttft = calculate_ttft(scope._timestamps, scope._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:
+ scope._llmresponse += content
+
+ if chunked.get('x_groq') is not None:
+ if chunked.get('x_groq').get('usage') is not None:
+ scope._input_tokens = chunked.get('x_groq').get('usage').get('prompt_tokens')
+ scope._output_tokens = chunked.get('x_groq').get('usage').get('completion_tokens')
+ scope._response_id = chunked.get('x_groq').get('id')
+ scope._response_model = chunked.get('x_groq').get('model')
+ scope._finish_reason = chunked.get('choices', [{}])[0].get('finish_reason')
+ scope._system_fingerprint = chunked.get('x_groq').get('system_fingerprint')
+ scope._end_time = time.time()
+
+
+def common_chat_logic(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content, disable_metrics, version, is_stream):
+ """
+ Process chat request and generate Telemetry
+ """
+
+ if len(scope._timestamps) > 1:
+ scope._tbt = calculate_tbt(scope._timestamps)
+
+ prompt = format_content(scope._kwargs.get("messages", []))
+ request_model = scope._kwargs.get("model", "mixtral-8x7b-32768")
+
+ cost = get_chat_model_cost(request_model, pricing_info, scope._input_tokens, scope._output_tokens)
+
+ # Common Span Attributes
+ common_span_attributes(scope,
+ SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT, SemanticConvention.GEN_AI_SYSTEM_GROQ,
+ scope._server_address, scope._server_port, request_model, scope._response_model,
+ environment, application_name, is_stream, scope._tbt, scope._ttft, version)
+
+ # Span Attributes for Request parameters
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SEED, scope._kwargs.get("seed", ""))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_FREQUENCY_PENALTY, scope._kwargs.get("frequency_penalty", 0.0))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MAX_TOKENS, scope._kwargs.get("max_completion_tokens", -1))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_PRESENCE_PENALTY, scope._kwargs.get("presence_penalty", 0.0))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_STOP_SEQUENCES, scope._kwargs.get("stop", []))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TEMPERATURE, scope._kwargs.get("temperature", 1.0))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_TOP_P, scope._kwargs.get("top_p", 1.0))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_USER, scope._kwargs.get("user", ""))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_SERVICE_TIER, scope._kwargs.get("service_tier", "on_demand"))
+
+ # Span Attributes for Response parameters
+ scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_ID, scope._response_id)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_FINISH_REASON, [scope._finish_reason])
+ scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_SYSTEM_FINGERPRINT, scope._system_fingerprint)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_OUTPUT_TYPE, "text" if isinstance(scope._llmresponse, str) else "json")
+
+ # Span Attributes for Cost and Tokens
+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_INPUT_TOKENS, scope._input_tokens)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_OUTPUT_TOKENS, scope._output_tokens)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_CLIENT_TOKEN_USAGE, scope._input_tokens + scope._output_tokens)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST, cost)
+
+ # Span Attributes for Tools
+ if scope._tools:
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_NAME, scope._tools.get("function", {}).get("name", ""))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_CALL_ID, str(scope._tools.get("id", "")))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_ARGS, str(scope._tools.get("function", {}).get("arguments", "")))
+
+ # Span Attributes for Content
+ if capture_message_content:
+ scope._span.set_attribute(SemanticConvention.GEN_AI_CONTENT_PROMPT, prompt)
+ scope._span.set_attribute(SemanticConvention.GEN_AI_CONTENT_COMPLETION, scope._llmresponse)
+
+ # To be removed one the change to span_attributes (from span events) is complete
+ scope._span.add_event(
+ name=SemanticConvention.GEN_AI_CONTENT_PROMPT_EVENT,
+ attributes={
+ SemanticConvention.GEN_AI_CONTENT_PROMPT: prompt,
+ },
+ )
+ scope._span.add_event(
+ name=SemanticConvention.GEN_AI_CONTENT_COMPLETION_EVENT,
+ attributes={
+ SemanticConvention.GEN_AI_CONTENT_COMPLETION: scope._llmresponse,
+ },
+ )
+
+ scope._span.set_status(Status(StatusCode.OK))
+
+ # Metrics
+ if not disable_metrics:
+ record_completion_metrics(metrics, SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT, SemanticConvention.GEN_AI_SYSTEM_GROQ,
+ scope._server_address, scope._server_port, request_model, scope._response_model, environment,
+ application_name, scope._start_time, scope._end_time, scope._input_tokens, scope._output_tokens,
+ cost, scope._tbt, scope._ttft)
+
+def process_streaming_chat_response(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content=False, disable_metrics=False, version=""):
+ """
+ Process chat request and generate Telemetry
+ """
+
+ common_chat_logic(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content, disable_metrics, version, is_stream=True)
+
+def process_chat_response(response, request_model, pricing_info, server_port, server_address,
+ environment, application_name, metrics, start_time, span, capture_message_content=False,
+ disable_metrics=False, version="1.0.0", **kwargs):
+ """
+ Process chat request and generate Telemetry
+ """
+
+ # Create scope object
+ scope = type("GenericScope", (), {})()
+ response_dict = response_as_dict(response)
+
+ scope._start_time = start_time
+ scope._end_time = time.time()
+ scope._span = span
+ scope._llmresponse = " ".join(
+ (choice.get("message", {}).get("content") or "")
+ for choice in response_dict.get("choices", [])
+ )
+ scope._response_id = response_dict.get("id")
+ scope._response_model = response_dict.get("model")
+ scope._input_tokens = response_dict.get("usage", {}).get("prompt_tokens", 0)
+ scope._output_tokens = response_dict.get("usage", {}).get("completion_tokens", 0)
+ scope._timestamps = []
+ scope._ttft, scope._tbt = scope._end_time - scope._start_time, 0
+ scope._server_address, scope._server_port = server_address, server_port
+ scope._kwargs = kwargs
+ scope._system_fingerprint = response_dict.get("system_fingerprint")
+ scope._finish_reason = str(response_dict.get("choices", [])[0].get("finish_reason", ""))
+
+ # Handle tool calls
+ if scope._kwargs.get("tools"):
+ scope._tools = response_dict.get("choices", [{}])[0].get("message", {}).get("tool_calls")
+ else:
+ scope._tools = None
+
+ common_chat_logic(scope, pricing_info, environment, application_name, metrics,
+ capture_message_content, disable_metrics, version, is_stream=False)
+
+ return response
diff --git a/sdk/python/src/openlit/instrumentation/ollama/async_ollama.py b/sdk/python/src/openlit/instrumentation/ollama/async_ollama.py
index 27b6fa46..5ce7de9b 100644
--- a/sdk/python/src/openlit/instrumentation/ollama/async_ollama.py
+++ b/sdk/python/src/openlit/instrumentation/ollama/async_ollama.py
@@ -42,7 +42,7 @@ def async_chat(version, environment, application_name,
self._llmresponse = ""
self._response_model = ""
self._finish_reason = ""
- self._tool_calls = []
+ self._tools = []
self._input_tokens = 0
self._output_tokens = 0
self._response_role = ""
@@ -102,7 +102,8 @@ def async_chat(version, environment, application_name,
streaming = kwargs.get("stream", False)
server_address, server_port = set_server_address_and_port(instance, "127.0.0.1", 11434)
- request_model = kwargs.get("model")
+ json_body = kwargs.get("json", {}) or {}
+ request_model = json_body.get("model") or kwargs.get("model")
span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
diff --git a/sdk/python/src/openlit/instrumentation/ollama/ollama.py b/sdk/python/src/openlit/instrumentation/ollama/ollama.py
index 1e7430a2..0b1a9ebe 100644
--- a/sdk/python/src/openlit/instrumentation/ollama/ollama.py
+++ b/sdk/python/src/openlit/instrumentation/ollama/ollama.py
@@ -42,7 +42,7 @@ def chat(version, environment, application_name,
self._llmresponse = ""
self._response_model = ""
self._finish_reason = ""
- self._tool_calls = []
+ self._tools = []
self._input_tokens = 0
self._output_tokens = 0
self._response_role = ""
@@ -102,7 +102,8 @@ def chat(version, environment, application_name,
streaming = kwargs.get("stream", False)
server_address, server_port = set_server_address_and_port(instance, "127.0.0.1", 11434)
- request_model = kwargs.get("model")
+ json_body = kwargs.get("json", {}) or {}
+ request_model = json_body.get("model") or kwargs.get("model")
span_name = f"{SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT} {request_model}"
diff --git a/sdk/python/src/openlit/instrumentation/ollama/utils.py b/sdk/python/src/openlit/instrumentation/ollama/utils.py
index 9f877ddb..7bf2dd16 100644
--- a/sdk/python/src/openlit/instrumentation/ollama/utils.py
+++ b/sdk/python/src/openlit/instrumentation/ollama/utils.py
@@ -57,7 +57,7 @@ def process_chunk(self, chunk):
self._llmresponse += chunked.get("message", {}).get("content", "")
if chunked.get("message", {}).get("tool_calls"):
- self._tool_calls = chunked["message"]["tool_calls"]
+ self._tools = chunked["message"]["tool_calls"]
if chunked.get("eval_count"):
self._response_role = chunked.get("message", {}).get("role", "")
@@ -138,10 +138,10 @@ def common_chat_logic(scope, gen_ai_endpoint, pricing_info, environment, applica
scope._span.set_attribute(SemanticConvention.GEN_AI_USAGE_COST, cost)
# Span Attributes for Tools
- if hasattr(scope, "_tool_calls") and scope._tool_calls:
- tool_call = scope._tool_calls[0]
- scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_NAME, tool_call.get("function", {}).get("name", ""))
- scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_ARGS, str(tool_call.get("function", {}).get("arguments", "")))
+ if scope._tools is not None:
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_NAME, scope._tools.get("function","")).get("name","")
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_CALL_ID, str(scope._tools.get("id","")))
+ scope._span.set_attribute(SemanticConvention.GEN_AI_TOOL_ARGS, str(scope._tools.get("function","").get("arguments","")))
# Span Attributes for Content
if capture_message_content:
@@ -244,7 +244,11 @@ def process_chat_response(response, gen_ai_endpoint, pricing_info, server_port,
scope._tbt = 0
scope._server_address, scope._server_port = server_address, server_port
scope._kwargs = kwargs
- scope._tool_calls = response_dict.get("message", {}).get("tool_calls", [])
+
+ if scope._kwargs.get("tools"):
+ scope._tools = response_dict.get("choices")[0].get("message").get("tool_calls")
+ else:
+ scope._tools = None
common_chat_logic(scope, gen_ai_endpoint, pricing_info, environment, application_name, metrics,
capture_message_content, disable_metrics, version)
diff --git a/sdk/python/src/openlit/instrumentation/premai/utils.py b/sdk/python/src/openlit/instrumentation/premai/utils.py
index 4a88ca30..17e7b7ac 100644
--- a/sdk/python/src/openlit/instrumentation/premai/utils.py
+++ b/sdk/python/src/openlit/instrumentation/premai/utils.py
@@ -3,7 +3,6 @@ PremAI OpenTelemetry instrumentation utility functions
"""
import time
-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT
from opentelemetry.trace import Status, StatusCode
from openlit.__helpers import (
@@ -13,7 +12,9 @@ from openlit.__helpers import (
get_chat_model_cost,
get_embed_model_cost,
general_tokens,
- create_metrics_attributes,
+ common_span_attributes,
+ record_completion_metrics,
+ record_embedding_metrics,
)
from openlit.semcov import SemanticConvention
@@ -66,77 +67,6 @@ def process_chunk(scope, chunk):
scope._response_model = chunked.get("model")
scope._end_time = time.time()
-def common_span_attributes(scope, gen_ai_operation, gen_ai_system, server_address, server_port,
- request_model, response_model, environment, application_name, is_stream, tbt, ttft, version):
- """
- Set common span attributes for both chat and RAG operations.
- """
-
- scope._span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
- scope._span.set_attribute(SemanticConvention.GEN_AI_OPERATION, gen_ai_operation)
- scope._span.set_attribute(SemanticConvention.GEN_AI_SYSTEM, gen_ai_system)
- scope._span.set_attribute(SemanticConvention.SERVER_ADDRESS, server_address)
- scope._span.set_attribute(SemanticConvention.SERVER_PORT, server_port)
- scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL, request_model)
- scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL, response_model)
- scope._span.set_attribute(DEPLOYMENT_ENVIRONMENT, environment)
- scope._span.set_attribute(SERVICE_NAME, application_name)
- scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM, is_stream)
- scope._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TBT, tbt)
- scope._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT, ttft)
- scope._span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION, version)
-
-def record_completion_metrics(metrics, gen_ai_operation, gen_ai_system, server_address, server_port,
- request_model, response_model, environment, application_name, start_time, end_time,
- input_tokens, output_tokens, cost, tbt=None, ttft=None):
- """
- Record completion-specific metrics for the operation.
- """
-
- attributes = create_metrics_attributes(
- operation=gen_ai_operation,
- system=gen_ai_system,
- server_address=server_address,
- server_port=server_port,
- request_model=request_model,
- response_model=response_model,
- service_name=application_name,
- deployment_environment=environment,
- )
- metrics["genai_client_operation_duration"].record(end_time - start_time, attributes)
- metrics["genai_requests"].add(1, attributes)
- metrics["genai_prompt_tokens"].add(input_tokens, attributes)
- metrics["genai_completion_tokens"].add(output_tokens, attributes)
- metrics["genai_client_usage_tokens"].record(input_tokens + output_tokens, attributes)
- metrics["genai_cost"].record(cost, attributes)
- if tbt is not None:
- metrics["genai_server_tbt"].record(tbt, attributes)
- if ttft is not None:
- metrics["genai_server_ttft"].record(ttft, attributes)
-
-def record_embedding_metrics(metrics, gen_ai_operation, gen_ai_system, server_address, server_port,
- request_model, response_model, environment, application_name, start_time, end_time,
- input_tokens, cost):
- """
- Record embedding-specific metrics for the operation.
- """
-
- attributes = create_metrics_attributes(
- operation=gen_ai_operation,
- system=gen_ai_system,
- server_address=server_address,
- server_port=server_port,
- request_model=request_model,
- response_model=response_model,
- service_name=application_name,
- deployment_environment=environment,
- )
- 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_prompt_tokens"].add(input_tokens, attributes)
- metrics["genai_cost"].record(cost, attributes)
-
def common_chat_logic(scope, pricing_info, environment, application_name, metrics,
capture_message_content, disable_metrics, version, is_stream):
"""
diff --git a/sdk/python/src/openlit/instrumentation/reka/utils.py b/sdk/python/src/openlit/instrumentation/reka/utils.py
index 0f46c7db..e6a69986 100644
--- a/sdk/python/src/openlit/instrumentation/reka/utils.py
+++ b/sdk/python/src/openlit/instrumentation/reka/utils.py
@@ -3,14 +3,14 @@ Reka OpenTelemetry instrumentation utility functions
"""
import time
-from opentelemetry.sdk.resources import SERVICE_NAME, TELEMETRY_SDK_NAME, DEPLOYMENT_ENVIRONMENT
from opentelemetry.trace import Status, StatusCode
from openlit.__helpers import (
response_as_dict,
calculate_tbt,
get_chat_model_cost,
- create_metrics_attributes,
+ common_span_attributes,
+ record_completion_metrics,
)
from openlit.semcov import SemanticConvention
@@ -36,54 +36,6 @@ def format_content(messages):
return "\n".join(formatted_messages)
-def common_span_attributes(scope, gen_ai_operation, gen_ai_system, server_address, server_port,
- request_model, response_model, environment, application_name, is_stream, tbt, ttft, version):
- """
- Set common span attributes for both chat and RAG operations.
- """
-
- scope._span.set_attribute(TELEMETRY_SDK_NAME, "openlit")
- scope._span.set_attribute(SemanticConvention.GEN_AI_OPERATION, gen_ai_operation)
- scope._span.set_attribute(SemanticConvention.GEN_AI_SYSTEM, gen_ai_system)
- scope._span.set_attribute(SemanticConvention.SERVER_ADDRESS, server_address)
- scope._span.set_attribute(SemanticConvention.SERVER_PORT, server_port)
- scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_MODEL, request_model)
- scope._span.set_attribute(SemanticConvention.GEN_AI_RESPONSE_MODEL, response_model)
- scope._span.set_attribute(DEPLOYMENT_ENVIRONMENT, environment)
- scope._span.set_attribute(SERVICE_NAME, application_name)
- scope._span.set_attribute(SemanticConvention.GEN_AI_REQUEST_IS_STREAM, is_stream)
- scope._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TBT, tbt)
- scope._span.set_attribute(SemanticConvention.GEN_AI_SERVER_TTFT, ttft)
- scope._span.set_attribute(SemanticConvention.GEN_AI_SDK_VERSION, version)
-
-def record_common_metrics(metrics, gen_ai_operation, gen_ai_system, server_address, server_port,
- request_model, response_model, environment, application_name, start_time, end_time,
- input_tokens, output_tokens, cost, tbt=None, ttft=None):
- """
- Record common metrics for the operation.
- """
-
- attributes = create_metrics_attributes(
- operation=gen_ai_operation,
- system=gen_ai_system,
- server_address=server_address,
- server_port=server_port,
- request_model=request_model,
- response_model=response_model,
- service_name=application_name,
- deployment_environment=environment,
- )
- metrics["genai_client_operation_duration"].record(end_time - start_time, attributes)
- metrics["genai_requests"].add(1, attributes)
- metrics["genai_prompt_tokens"].add(input_tokens, attributes)
- metrics["genai_completion_tokens"].add(output_tokens, attributes)
- metrics["genai_client_usage_tokens"].record(input_tokens + output_tokens, attributes)
- metrics["genai_cost"].record(cost, attributes)
- if tbt is not None:
- metrics["genai_server_tbt"].record(tbt, attributes)
- if ttft is not None:
- metrics["genai_server_ttft"].record(ttft, attributes)
-
def common_chat_logic(scope, pricing_info, environment, application_name, metrics,
capture_message_content, disable_metrics, version, is_stream):
"""
@@ -151,7 +103,7 @@ def common_chat_logic(scope, pricing_info, environment, application_name, metric
# Metrics
if not disable_metrics:
- record_common_metrics(metrics, SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT, SemanticConvention.GEN_AI_SYSTEM_REKAAI,
+ record_completion_metrics(metrics, SemanticConvention.GEN_AI_OPERATION_TYPE_CHAT, SemanticConvention.GEN_AI_SYSTEM_REKAAI,
scope._server_address, scope._server_port, request_model, scope._response_model, environment,
application_name, scope._start_time, scope._end_time, scope._input_tokens, scope._output_tokens,
cost, scope._tbt, scope._ttft)