View Source OpenTelemetry.SemConv.Incubating.GenAiAttributes (OpenTelemetry.SemConv v1.27.0)

OpenTelemetry Semantic Conventions for Gen_Ai attributes.

Summary

Types

gen_ai_operation_name_values()

The name of the operation being performed.

gen_ai_system_values()

The Generative AI product as identified by the client or server instrumentation.

gen_ai_token_type_values()

The type of token being counted.

Functions

gen_ai_completion()

The full response received from the GenAI model.

gen_ai_operation_name()

The name of the operation being performed.

gen_ai_operation_name_values()
gen_ai_prompt()

The full prompt sent to the GenAI model.

gen_ai_request_frequency_penalty()

The frequency penalty setting for the GenAI request.

gen_ai_request_max_tokens()

The maximum number of tokens the model generates for a request.

gen_ai_request_model()

The name of the GenAI model a request is being made to.

gen_ai_request_presence_penalty()

The presence penalty setting for the GenAI request.

gen_ai_request_stop_sequences()

List of sequences that the model will use to stop generating further tokens.

gen_ai_request_temperature()

The temperature setting for the GenAI request.

gen_ai_request_top_k()

The top_k sampling setting for the GenAI request.

gen_ai_request_top_p()

The top_p sampling setting for the GenAI request.

gen_ai_response_finish_reasons()

Array of reasons the model stopped generating tokens, corresponding to each generation received.

gen_ai_response_id()

The unique identifier for the completion.

gen_ai_response_model()

The name of the model that generated the response.

gen_ai_system()

The Generative AI product as identified by the client or server instrumentation.

gen_ai_system_values()
gen_ai_token_type()

The type of token being counted.

gen_ai_token_type_values()
gen_ai_usage_completion_tokens() deprecated
gen_ai_usage_input_tokens()

The number of tokens used in the GenAI input (prompt).

gen_ai_usage_output_tokens()

The number of tokens used in the GenAI response (completion).

gen_ai_usage_prompt_tokens() deprecated

Types

Link to this type

gen_ai_operation_name_values()

View Source 
@type gen_ai_operation_name_values() :: %{
  chat: :chat,
  text_completion: :text_completion
}

The name of the operation being performed.

Enum Values

Link to this type

gen_ai_system_values()

View Source 
@type gen_ai_system_values() :: %{
  openai: :openai,
  vertex_ai: :vertex_ai,
  anthropic: :anthropic,
  cohere: :cohere
}

The Generative AI product as identified by the client or server instrumentation.

Enum Values

  • :openai e - OpenAI
  • :vertex_ai e - Vertex AI
  • :anthropic e - Anthropic
  • :cohere e - Cohere
Link to this type

gen_ai_token_type_values()

View Source 
@type gen_ai_token_type_values() :: %{input: :input, completion: :output}

The type of token being counted.

Enum Values

  • :input e - Input tokens (prompt, input, etc.)
  • :completion e - Output tokens (completion, response, etc.)

Functions

Link to this function

gen_ai_completion()

View Source 
@spec gen_ai_completion() :: :"gen_ai.completion"

The full response received from the GenAI model.

Value type

Value must be of type atom() | String.t().

Notes

It's RECOMMENDED to format completions as JSON string matching OpenAI messages format

Examples 

["[{'role': 'assistant', 'content': 'The capital of France is Paris.'}]"]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_completion()
:"gen_ai.completion"
Link to this function

gen_ai_operation_name()

View Source 
@spec gen_ai_operation_name() :: :"gen_ai.operation.name"

The name of the operation being performed.

Notes

If one of the predefined values applies, but specific system uses a different name it's RECOMMENDED to document it in the semantic conventions for specific GenAI system and use system-specific name in the instrumentation. If a different name is not documented, instrumentation libraries SHOULD use applicable predefined value.

iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_operation_name()
:"gen_ai.operation.name"

iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_operation_name_values().chat
:chat

iex> %{OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_operation_name() => OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_operation_name_values().chat}
%{:"gen_ai.operation.name" => :chat}
Link to this function

gen_ai_operation_name_values()

View Source 
@spec gen_ai_operation_name_values() :: gen_ai_operation_name_values()
Link to this function

gen_ai_prompt()

View Source 
@spec gen_ai_prompt() :: :"gen_ai.prompt"

The full prompt sent to the GenAI model.

Value type

Value must be of type atom() | String.t().

Notes

It's RECOMMENDED to format prompts as JSON string matching OpenAI messages format

Examples 

["[{'role': 'user', 'content': 'What is the capital of France?'}]"]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_prompt()
:"gen_ai.prompt"
Link to this function

gen_ai_request_frequency_penalty()

View Source 
@spec gen_ai_request_frequency_penalty() :: :"gen_ai.request.frequency_penalty"

The frequency penalty setting for the GenAI request.

Value type

Value must be of type float().

Examples 

[0.1]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_request_frequency_penalty()
:"gen_ai.request.frequency_penalty"
Link to this function

gen_ai_request_max_tokens()

View Source 
@spec gen_ai_request_max_tokens() :: :"gen_ai.request.max_tokens"

The maximum number of tokens the model generates for a request.

Value type

Value must be of type integer().

Examples 

[100]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_request_max_tokens()
:"gen_ai.request.max_tokens"
Link to this function

gen_ai_request_model()

View Source 
@spec gen_ai_request_model() :: :"gen_ai.request.model"

The name of the GenAI model a request is being made to.

Value type

Value must be of type atom() | String.t().

Examples 

gpt-4
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_request_model()
:"gen_ai.request.model"
Link to this function

gen_ai_request_presence_penalty()

View Source 
@spec gen_ai_request_presence_penalty() :: :"gen_ai.request.presence_penalty"

The presence penalty setting for the GenAI request.

Value type

Value must be of type float().

Examples 

[0.1]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_request_presence_penalty()
:"gen_ai.request.presence_penalty"
Link to this function

gen_ai_request_stop_sequences()

View Source 
@spec gen_ai_request_stop_sequences() :: :"gen_ai.request.stop_sequences"

List of sequences that the model will use to stop generating further tokens.

Value type

Value must be of type [atom() | String.t()].

Examples 

["forest", "lived"]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_request_stop_sequences()
:"gen_ai.request.stop_sequences"
Link to this function

gen_ai_request_temperature()

View Source 
@spec gen_ai_request_temperature() :: :"gen_ai.request.temperature"

The temperature setting for the GenAI request.

Value type

Value must be of type float().

Examples 

[0.0]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_request_temperature()
:"gen_ai.request.temperature"
Link to this function

gen_ai_request_top_k()

View Source 
@spec gen_ai_request_top_k() :: :"gen_ai.request.top_k"

The top_k sampling setting for the GenAI request.

Value type

Value must be of type float().

Examples 

[1.0]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_request_top_k()
:"gen_ai.request.top_k"
Link to this function

gen_ai_request_top_p()

View Source 
@spec gen_ai_request_top_p() :: :"gen_ai.request.top_p"

The top_p sampling setting for the GenAI request.

Value type

Value must be of type float().

Examples 

[1.0]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_request_top_p()
:"gen_ai.request.top_p"
Link to this function

gen_ai_response_finish_reasons()

View Source 
@spec gen_ai_response_finish_reasons() :: :"gen_ai.response.finish_reasons"

Array of reasons the model stopped generating tokens, corresponding to each generation received.

Value type

Value must be of type [atom() | String.t()].

Examples 

["stop"]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_response_finish_reasons()
:"gen_ai.response.finish_reasons"
Link to this function

gen_ai_response_id()

View Source 
@spec gen_ai_response_id() :: :"gen_ai.response.id"

The unique identifier for the completion.

Value type

Value must be of type atom() | String.t().

Examples 

["chatcmpl-123"]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_response_id()
:"gen_ai.response.id"
Link to this function

gen_ai_response_model()

View Source 
@spec gen_ai_response_model() :: :"gen_ai.response.model"

The name of the model that generated the response.

Value type

Value must be of type atom() | String.t().

Examples 

["gpt-4-0613"]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_response_model()
:"gen_ai.response.model"
Link to this function

gen_ai_system()

View Source 
@spec gen_ai_system() :: :"gen_ai.system"

The Generative AI product as identified by the client or server instrumentation.

Notes

The gen_ai.system describes a family of GenAI models with specific model identified
by gen_ai.request.model and gen_ai.response.model attributes.

The actual GenAI product may differ from the one identified by the client.
For example, when using OpenAI client libraries to communicate with Mistral, the gen_ai.system
is set to openai based on the instrumentation's best knowledge.

For custom model, a custom friendly name SHOULD be used.
If none of these options apply, the gen_ai.system SHOULD be set to _OTHER.

Examples 

openai
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_system()
:"gen_ai.system"

iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_system_values().openai
:openai

iex> %{OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_system() => OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_system_values().openai}
%{:"gen_ai.system" => :openai}
Link to this function

gen_ai_system_values()

View Source 
@spec gen_ai_system_values() :: gen_ai_system_values()
Link to this function

gen_ai_token_type()

View Source 
@spec gen_ai_token_type() :: :"gen_ai.token.type"

The type of token being counted.

Examples 

["input", "output"]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_token_type()
:"gen_ai.token.type"

iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_token_type_values().input
:input

iex> %{OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_token_type() => OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_token_type_values().input}
%{:"gen_ai.token.type" => :input}
Link to this function

gen_ai_token_type_values()

View Source 
@spec gen_ai_token_type_values() :: gen_ai_token_type_values()
Link to this function

gen_ai_usage_completion_tokens()

View Source
This function is deprecated. Replaced by `gen_ai.usage.output_tokens` attribute. . 
@spec gen_ai_usage_completion_tokens() :: :"gen_ai.usage.completion_tokens"
Link to this function

gen_ai_usage_input_tokens()

View Source 
@spec gen_ai_usage_input_tokens() :: :"gen_ai.usage.input_tokens"

The number of tokens used in the GenAI input (prompt).

Value type

Value must be of type integer().

Examples 

[100]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_usage_input_tokens()
:"gen_ai.usage.input_tokens"
Link to this function

gen_ai_usage_output_tokens()

View Source 
@spec gen_ai_usage_output_tokens() :: :"gen_ai.usage.output_tokens"

The number of tokens used in the GenAI response (completion).

Value type

Value must be of type integer().

Examples 

[180]
iex> OpenTelemetry.SemConv.Incubating.GenAiAttributes.gen_ai_usage_output_tokens()
:"gen_ai.usage.output_tokens"
Link to this function

gen_ai_usage_prompt_tokens()

View Source
This function is deprecated. Replaced by `gen_ai.usage.input_tokens` attribute. . 
@spec gen_ai_usage_prompt_tokens() :: :"gen_ai.usage.prompt_tokens"