Ignore Specific Error Classes in Auto-Instrumenting Langchain Traces

Would there be a way to ignore certain error classes within traces when auto-instrumenting langchain? Odd request, I know, but unfortunately langgraph uses exception handling for control flow, which means that every successful trace during a multi-graph execution will show an error response within phoenix. Here’s a link showing this behavior when it was introduced. https://github.com/langchain-ai/langgraph/pull/2520/files#diff-60253f1ec99c8d9b4e1b4fec9573d904c7065d6b0942e1143787b73d68d80d4cR71-R73