@zen-agentchecks…zen/fix-oversized-llm-calls-dgx91e → master1 files · +59 −0updated 1mo agoThe learning pipeline was failing on ~28 sessions per day due to oversized prompts sent to OpenAI's GPT-5.2 API:
These failures waste API quota and generate noise in logs. The root cause is that flex_chat_completions_parse() had no pre-flight size checks — it sent every request regardless of size, and the errors were only caught after the API rejected them.
Fix: Added three pre-flight checks to flex_chat_completions_parse():
These checks log a warning and return None early, matching the existing error-handling pattern. Sessions that are too large are simply skipped (workflow-only results are still saved when error analysis is skipped).
ruff check and ruff format --check — all passingOrigin: cron-5d55c321e47a / zen-cron-7bf66f8ab915 Triggered by: dhawal@composio.dev
Addressed Bugbot comment: added lone surrogates (\ud800-\udfff) to the sanitization regex in e42364e. The regex now correctly strips C0 controls, C1 controls, and lone surrogates as documented.