comment about caching

getsentry · Feb 6, 2025 · 471d807 · 471d807
1 parent 3b13a33
commit 471d807
Showing 1 changed file with 2 additions and 0 deletions.
diff --git a/src/seer/automation/codegen/relevant_warnings_component.py b/src/seer/automation/codegen/relevant_warnings_component.py
@@ -106,6 +106,8 @@ def invoke(
         # TODO: handle LLM API errors in this loop by moving on
         relevant_warning_results: list[RelevantWarningResult] = []
         candidate_associations = request.candidate_associations[: self._max_associations_analyzed]
+        # The time limit for OpenAI prompt caching is 5-10 minutes, so no point in sorting by
+        # issue.group_id
         for warning, issue in tqdm(candidate_associations, desc="Predicting relevance"):
             completion = llm_client.generate_structured(
                 model=OpenAiProvider.model("gpt-4o-mini-2024-07-18"),