Problem
OTel GenAI semconv recommends capturing gen_ai.request.stop_sequences (string[]). Google's GenerateContentConfig has a stop_sequences parameter, but it is not captured by the integration.
Context
Other config parameters (temperature, top_p, top_k, max_output_tokens, presence_penalty, frequency_penalty, seed) are already captured in the param loop at utils.py:834-846. Adding stop_sequences follows the same pattern.
Files
sentry_sdk/integrations/google_genai/utils.py — set_span_data_for_request()
sentry_sdk/consts.py — may need a new SPANDATA constant if one doesn't exist
Acceptance Criteria
stop_sequences from GenerateContentConfig is captured on the span
- Test added to verify the attribute is set when stop_sequences is provided
Problem
OTel GenAI semconv recommends capturing
gen_ai.request.stop_sequences(string[]). Google'sGenerateContentConfighas astop_sequencesparameter, but it is not captured by the integration.Context
Other config parameters (temperature, top_p, top_k, max_output_tokens, presence_penalty, frequency_penalty, seed) are already captured in the param loop at
utils.py:834-846. Addingstop_sequencesfollows the same pattern.Files
sentry_sdk/integrations/google_genai/utils.py—set_span_data_for_request()sentry_sdk/consts.py— may need a newSPANDATAconstant if one doesn't existAcceptance Criteria
stop_sequencesfromGenerateContentConfigis captured on the span