Data sources are used by AI agents and RAG applications to store grounding data. A data source may be an external database, object store, document collection, website, or any other storage system used by the GenAI agent or application. The @code gen_ai.data_source.id @endcode SHOULD match the identifier used by the GenAI system rather than a name specific to the external storage, such as a database or object store. Semantic conventions referencing @code gen_ai.data_source.id @endcode MAY also leverage additional attributes, such as @code db.* @endcode, to further identify and describe the data source. */ static constexpr const char *kGenAiDataSourceId = "gen_ai.data_source.id"; /** The number of dimensions the resulting output embeddings should have. */ static constexpr const char *kGenAiEmbeddingsDimensionCount = "gen_ai.embeddings.dimension.count"; /** A free-form explanation for the assigned score provided by the evaluator. */ static constexpr const char *kGenAiEvaluationExplanation = "gen_ai.evaluation.explanation"; /** The name of the evaluation metric used for the GenAI response. */ static constexpr const char *kGenAiEvaluationName = "gen_ai.evaluation.name"; /** Human readable label for evaluation.
This attribute provides a human-readable interpretation of the evaluation score produced by an evaluator. For example, a score value of 1 could mean "relevant" in one evaluation system and "not relevant" in another, depending on the scoring range and evaluator. The label SHOULD have low cardinality. Possible values depend on the evaluation metric and evaluator used; implementations SHOULD document the possible values. */ static constexpr const char *kGenAiEvaluationScoreLabel = "gen_ai.evaluation.score.label"; /** The evaluation score returned by the evaluator. */ static constexpr const char *kGenAiEvaluationScoreValue = "gen_ai.evaluation.score.value"; /** The chat history provided to the model as an input.
Instrumentations MUST follow Input messages JSON schema. When the attribute is recorded on events, it MUST be recorded in structured form. When recorded on spans, it MAY be recorded as a JSON string if structured format is not supported and SHOULD be recorded in structured form otherwise.
Messages MUST be provided in the order they were sent to the model. Instrumentations MAY provide a way for users to filter or truncate input messages.
[!Warning] This attribute is likely to contain sensitive information including user/PII data.
See Recording content on attributes section for more details. */ static constexpr const char *kGenAiInputMessages = "gen_ai.input.messages"; /** Deprecated, use @code gen_ai.output.type @endcode. @deprecated {"note": "Replaced by @code gen_ai.output.type @endcode.", "reason": "renamed", "renamed_to": "gen_ai.output.type"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiOpenaiRequestResponseFormat = "gen_ai.openai.request.response_format"; /** Deprecated, use @code gen_ai.request.seed @endcode. @deprecated {"note": "Replaced by @code gen_ai.request.seed @endcode.", "reason": "renamed", "renamed_to": "gen_ai.request.seed"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiOpenaiRequestSeed = "gen_ai.openai.request.seed"; /** Deprecated, use @code openai.request.service_tier @endcode. @deprecated {"note": "Replaced by @code openai.request.service_tier @endcode.", "reason": "renamed", "renamed_to": "openai.request.service_tier"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiOpenaiRequestServiceTier = "gen_ai.openai.request.service_tier"; /** Deprecated, use @code openai.response.service_tier @endcode. @deprecated {"note": "Replaced by @code openai.response.service_tier @endcode.", "reason": "renamed", "renamed_to": "openai.response.service_tier"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiOpenaiResponseServiceTier = "gen_ai.openai.response.service_tier"; /** Deprecated, use @code openai.response.system_fingerprint @endcode. @deprecated {"note": "Replaced by @code openai.response.system_fingerprint @endcode.", "reason": "renamed", "renamed_to": "openai.response.system_fingerprint"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiOpenaiResponseSystemFingerprint = "gen_ai.openai.response.system_fingerprint"; /** The name of the operation being performed.
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. */ static constexpr const char *kGenAiOperationName = "gen_ai.operation.name"; /** Messages returned by the model where each message represents a specific model response (choice, candidate).
Instrumentations MUST follow Output messages JSON schema
Each message represents a single output choice/candidate generated by the model. Each message corresponds to exactly one generation (choice/candidate) and vice versa - one choice cannot be split across multiple messages or one message cannot contain parts from multiple choices.
When the attribute is recorded on events, it MUST be recorded in structured form. When recorded on spans, it MAY be recorded as a JSON string if structured format is not supported and SHOULD be recorded in structured form otherwise.
Instrumentations MAY provide a way for users to filter or truncate output messages.
[!Warning] This attribute is likely to contain sensitive information including user/PII data.
See Recording content on attributes section for more details. */ static constexpr const char *kGenAiOutputMessages = "gen_ai.output.messages"; /** Represents the content type requested by the client.
This attribute SHOULD be used when the client requests output of a specific type. The model may return zero or more outputs of this type. This attribute specifies the output modality and not the actual output format. For example, if an image is requested, the actual output could be a URL pointing to an image file. Additional output format details may be recorded in the future in the @code gen_ai.output.{type}.* @endcode attributes. */ static constexpr const char *kGenAiOutputType = "gen_ai.output.type"; /** Deprecated, use Event API to report prompt contents. @deprecated {"note": "Removed, no replacement at this time.", "reason": "obsoleted"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiPrompt = "gen_ai.prompt"; /** The name of the prompt that uniquely identifies it. */ static constexpr const char *kGenAiPromptName = "gen_ai.prompt.name"; /** The Generative AI provider as identified by the client or server instrumentation.
The attribute SHOULD be set based on the instrumentation's best knowledge and may differ from the actual model provider.
Multiple providers, including Azure OpenAI, Gemini, and AI hosting platforms are accessible using the OpenAI REST API and corresponding client libraries, but may proxy or host models from different providers.
The @code gen_ai.request.model @endcode, @code gen_ai.response.model @endcode, and @code server.address @endcode attributes may help identify the actual system in use.
The @code gen_ai.provider.name @endcode attribute acts as a discriminator that identifies the GenAI telemetry format flavor specific to that provider within GenAI semantic conventions. It SHOULD be set consistently with provider-specific attributes and signals. For example, GenAI spans, metrics, and events related to AWS Bedrock should have the @code gen_ai.provider.name @endcode set to @code aws.bedrock @endcode and include applicable @code aws.bedrock.* @endcode attributes and are not expected to include @code openai.* @endcode attributes. */ static constexpr const char *kGenAiProviderName = "gen_ai.provider.name"; /** The target number of candidate completions to return. */ static constexpr const char *kGenAiRequestChoiceCount = "gen_ai.request.choice.count"; /** The encoding formats requested in an embeddings operation, if specified.
In some GenAI systems the encoding formats are called embedding types. Also, some GenAI systems only accept a single format per request. */ static constexpr const char *kGenAiRequestEncodingFormats = "gen_ai.request.encoding_formats"; /** The frequency penalty setting for the GenAI request. */ static constexpr const char *kGenAiRequestFrequencyPenalty = "gen_ai.request.frequency_penalty"; /** The maximum number of tokens the model generates for a request. */ static constexpr const char *kGenAiRequestMaxTokens = "gen_ai.request.max_tokens"; /** The name of the GenAI model a request is being made to. */ static constexpr const char *kGenAiRequestModel = "gen_ai.request.model"; /** The presence penalty setting for the GenAI request. */ static constexpr const char *kGenAiRequestPresencePenalty = "gen_ai.request.presence_penalty"; /** Requests with same seed value more likely to return same result. */ static constexpr const char *kGenAiRequestSeed = "gen_ai.request.seed"; /** List of sequences that the model will use to stop generating further tokens. */ static constexpr const char *kGenAiRequestStopSequences = "gen_ai.request.stop_sequences"; /** The temperature setting for the GenAI request. */ static constexpr const char *kGenAiRequestTemperature = "gen_ai.request.temperature"; /** The top_k sampling setting for the GenAI request. */ static constexpr const char *kGenAiRequestTopK = "gen_ai.request.top_k"; /** The top_p sampling setting for the GenAI request. */ static constexpr const char *kGenAiRequestTopP = "gen_ai.request.top_p"; /** Array of reasons the model stopped generating tokens, corresponding to each generation received. */ static constexpr const char *kGenAiResponseFinishReasons = "gen_ai.response.finish_reasons"; /** The unique identifier for the completion. */ static constexpr const char *kGenAiResponseId = "gen_ai.response.id"; /** The name of the model that generated the response. */ static constexpr const char *kGenAiResponseModel = "gen_ai.response.model"; /** The documents retrieved.
Instrumentations MUST follow Retrieval documents JSON schema. When the attribute is recorded on events, it MUST be recorded in structured form. When recorded on spans, it MAY be recorded as a JSON string if structured format is not supported and SHOULD be recorded in structured form otherwise.
Each document object SHOULD contain at least the following properties: @code id @endcode (string): A unique identifier for the document, @code score @endcode (double): The relevance score of the document */ static constexpr const char *kGenAiRetrievalDocuments = "gen_ai.retrieval.documents"; /** The query text used for retrieval.
[!Warning] This attribute may contain sensitive information.*/ static constexpr const char *kGenAiRetrievalQueryText = "gen_ai.retrieval.query.text"; /** Deprecated, use @code gen_ai.provider.name @endcode instead. @deprecated {"note": "Replaced by @code gen_ai.provider.name @endcode.", "reason": "renamed", "renamed_to": "gen_ai.provider.name"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiSystem = "gen_ai.system"; /** The system message or instructions provided to the GenAI model separately from the chat history.
This attribute SHOULD be used when the corresponding provider or API allows to provide system instructions or messages separately from the chat history.
Instructions that are part of the chat history SHOULD be recorded in @code gen_ai.input.messages @endcode attribute instead.
Instrumentations MUST follow System instructions JSON schema.
When recorded on spans, it MAY be recorded as a JSON string if structured format is not supported and SHOULD be recorded in structured form otherwise.
Instrumentations MAY provide a way for users to filter or truncate system instructions.
[!Warning] This attribute may contain sensitive information.
See Recording content on attributes section for more details. */ static constexpr const char *kGenAiSystemInstructions = "gen_ai.system_instructions"; /** The type of token being counted. */ static constexpr const char *kGenAiTokenType = "gen_ai.token.type"; /** Parameters passed to the tool call.
[!WARNING] This attribute may contain sensitive information.
It's expected to be an object - in case a serialized string is available to the instrumentation, the instrumentation SHOULD do the best effort to deserialize it to an object. When recorded on spans, it MAY be recorded as a JSON string if structured format is not supported and SHOULD be recorded in structured form otherwise. */ static constexpr const char *kGenAiToolCallArguments = "gen_ai.tool.call.arguments"; /** The tool call identifier. */ static constexpr const char *kGenAiToolCallId = "gen_ai.tool.call.id"; /** The result returned by the tool call (if any and if execution was successful).
[!WARNING] This attribute may contain sensitive information.
It's expected to be an object - in case a serialized string is available to the instrumentation, the instrumentation SHOULD do the best effort to deserialize it to an object. When recorded on spans, it MAY be recorded as a JSON string if structured format is not supported and SHOULD be recorded in structured form otherwise. */ static constexpr const char *kGenAiToolCallResult = "gen_ai.tool.call.result"; /** The list of source system tool definitions available to the GenAI agent or model.
The value of this attribute matches source system tool definition format.
It's expected to be an array of objects where each object represents a tool definition. In case a serialized string is available to the instrumentation, the instrumentation SHOULD do the best effort to deserialize it to an array. When recorded on spans, it MAY be recorded as a JSON string if structured format is not supported and SHOULD be recorded in structured form otherwise.
Since this attribute could be large, it's NOT RECOMMENDED to populate it by default. Instrumentations MAY provide a way to enable populating this attribute. */ static constexpr const char *kGenAiToolDefinitions = "gen_ai.tool.definitions"; /** The tool description. */ static constexpr const char *kGenAiToolDescription = "gen_ai.tool.description"; /** Name of the tool utilized by the agent. */ static constexpr const char *kGenAiToolName = "gen_ai.tool.name"; /** Type of the tool utilized by the agent
Extension: A tool executed on the agent-side to directly call external APIs, bridging the gap between the agent and real-world systems. Agent-side operations involve actions that are performed by the agent on the server or within the agent's controlled environment. Function: A tool executed on the client-side, where the agent generates parameters for a predefined function, and the client executes the logic. Client-side operations are actions taken on the user's end or within the client application. Datastore: A tool used by the agent to access and query structured or unstructured external data for retrieval-augmented tasks or knowledge updates. */ static constexpr const char *kGenAiToolType = "gen_ai.tool.type"; /** The number of input tokens written to a provider-managed cache.
The value SHOULD be included in @code gen_ai.usage.input_tokens @endcode. */ static constexpr const char *kGenAiUsageCacheCreationInputTokens = "gen_ai.usage.cache_creation.input_tokens"; /** The number of input tokens served from a provider-managed cache.
The value SHOULD be included in @code gen_ai.usage.input_tokens @endcode. */ static constexpr const char *kGenAiUsageCacheReadInputTokens = "gen_ai.usage.cache_read.input_tokens"; /** Deprecated, use @code gen_ai.usage.output_tokens @endcode instead. @deprecated {"note": "Replaced by @code gen_ai.usage.output_tokens @endcode.", "reason": "renamed", "renamed_to": "gen_ai.usage.output_tokens"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiUsageCompletionTokens = "gen_ai.usage.completion_tokens"; /** The number of tokens used in the GenAI input (prompt).
This value SHOULD include all types of input tokens, including cached tokens. Instrumentations SHOULD make a best effort to populate this value, using a total provided by the provider when available or, depending on the provider API, by summing different token types parsed from the provider output. */ static constexpr const char *kGenAiUsageInputTokens = "gen_ai.usage.input_tokens"; /** The number of tokens used in the GenAI response (completion). */ static constexpr const char *kGenAiUsageOutputTokens = "gen_ai.usage.output_tokens"; /** Deprecated, use @code gen_ai.usage.input_tokens @endcode instead. @deprecated {"note": "Replaced by @code gen_ai.usage.input_tokens @endcode.", "reason": "renamed", "renamed_to": "gen_ai.usage.input_tokens"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGenAiUsagePromptTokens = "gen_ai.usage.prompt_tokens"; namespace GenAiOpenaiRequestResponseFormatValues { /** Text response format */ static constexpr const char *kText = "text"; /** JSON object response format */ static constexpr const char *kJsonObject = "json_object"; /** JSON schema response format */ static constexpr const char *kJsonSchema = "json_schema"; } // namespace GenAiOpenaiRequestResponseFormatValues namespace GenAiOpenaiRequestServiceTierValues { /** The system will utilize scale tier credits until they are exhausted. */ static constexpr const char *kAuto = "auto"; /** The system will utilize the default scale tier. */ static constexpr const char *kDefault = "default"; } // namespace GenAiOpenaiRequestServiceTierValues namespace GenAiOperationNameValues { /** Chat completion operation such as OpenAI Chat API */ static constexpr const char *kChat = "chat"; /** Multimodal content generation operation such as Gemini Generate Content */ static constexpr const char *kGenerateContent = "generate_content"; /** Text completions operation such as OpenAI Completions API (Legacy) */ static constexpr const char *kTextCompletion = "text_completion"; /** Embeddings operation such as OpenAI Create embeddings API */ static constexpr const char *kEmbeddings = "embeddings"; /** Retrieval operation such as OpenAI Search Vector Store API */ static constexpr const char *kRetrieval = "retrieval"; /** Create GenAI agent */ static constexpr const char *kCreateAgent = "create_agent"; /** Invoke GenAI agent */ static constexpr const char *kInvokeAgent = "invoke_agent"; /** Execute a tool */ static constexpr const char *kExecuteTool = "execute_tool"; } // namespace GenAiOperationNameValues namespace GenAiOutputTypeValues { /** Plain text */ static constexpr const char *kText = "text"; /** JSON object with known or unknown schema */ static constexpr const char *kJson = "json"; /** Image */ static constexpr const char *kImage = "image"; /** Speech */ static constexpr const char *kSpeech = "speech"; } // namespace GenAiOutputTypeValues namespace GenAiProviderNameValues { /** OpenAI */ static constexpr const char *kOpenai = "openai"; /** Any Google generative AI endpoint */ static constexpr const char *kGcpGenAi = "gcp.gen_ai"; /** Vertex AI */ static constexpr const char *kGcpVertexAi = "gcp.vertex_ai"; /** Gemini */ static constexpr const char *kGcpGemini = "gcp.gemini"; /** Anthropic */ static constexpr const char *kAnthropic = "anthropic"; /** Cohere */ static constexpr const char *kCohere = "cohere"; /** Azure AI Inference */ static constexpr const char *kAzureAiInference = "azure.ai.inference"; /** Azure OpenAI */ static constexpr const char *kAzureAiOpenai = "azure.ai.openai"; /** IBM Watsonx AI */ static constexpr const char *kIbmWatsonxAi = "ibm.watsonx.ai"; /** AWS Bedrock */ static constexpr const char *kAwsBedrock = "aws.bedrock"; /** Perplexity */ static constexpr const char *kPerplexity = "perplexity"; /** xAI */ static constexpr const char *kXAi = "x_ai"; /** DeepSeek */ static constexpr const char *kDeepseek = "deepseek"; /** Groq */ static constexpr const char *kGroq = "groq"; /** Mistral AI */ static constexpr const char *kMistralAi = "mistral_ai"; } // namespace GenAiProviderNameValues namespace GenAiSystemValues { /** OpenAI */ static constexpr const char *kOpenai = "openai"; /** Any Google generative AI endpoint */ static constexpr const char *kGcpGenAi = "gcp.gen_ai"; /** Vertex AI */ static constexpr const char *kGcpVertexAi = "gcp.vertex_ai"; /** Gemini */ static constexpr const char *kGcpGemini = "gcp.gemini"; /** Vertex AI @deprecated {"note": "Replaced by @code gcp.vertex_ai @endcode.", "reason": "renamed", "renamed_to": "gcp.vertex_ai"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kVertexAi = "vertex_ai"; /** Gemini @deprecated {"note": "Replaced by @code gcp.gemini @endcode.", "reason": "renamed", "renamed_to": "gcp.gemini"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kGemini = "gemini"; /** Anthropic */ static constexpr const char *kAnthropic = "anthropic"; /** Cohere */ static constexpr const char *kCohere = "cohere"; /** Azure AI Inference @deprecated {"note": "Replaced by @code azure.ai.inference @endcode.", "reason": "renamed", "renamed_to": "azure.ai.inference"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kAzAiInference = "az.ai.inference"; /** Azure OpenAI @deprecated {"note": "Replaced by @code azure.ai.openai @endcode.", "reason": "renamed", "renamed_to": "azure.ai.openai"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kAzAiOpenai = "az.ai.openai"; /** Azure AI Inference */ static constexpr const char *kAzureAiInference = "azure.ai.inference"; /** Azure OpenAI */ static constexpr const char *kAzureAiOpenai = "azure.ai.openai"; /** IBM Watsonx AI */ static constexpr const char *kIbmWatsonxAi = "ibm.watsonx.ai"; /** AWS Bedrock */ static constexpr const char *kAwsBedrock = "aws.bedrock"; /** Perplexity */ static constexpr const char *kPerplexity = "perplexity"; /** xAI */ static constexpr const char *kXai = "xai"; /** DeepSeek */ static constexpr const char *kDeepseek = "deepseek"; /** Groq */ static constexpr const char *kGroq = "groq"; /** Mistral AI */ static constexpr const char *kMistralAi = "mistral_ai"; } // namespace GenAiSystemValues namespace GenAiTokenTypeValues { /** Input tokens (prompt, input, etc.) */ static constexpr const char *kInput = "input"; /** Output tokens (completion, response, etc.) @deprecated {"note": "Replaced by @code output @endcode.", "reason": "renamed", "renamed_to": "output"} */ OPENTELEMETRY_DEPRECATED static constexpr const char *kCompletion = "output"; /** Output tokens (completion, response, etc.) */ static constexpr const char *kOutput = "output"; } // namespace GenAiTokenTypeValues } // namespace gen_ai } // namespace semconv OPENTELEMETRY_END_NAMESPACE