Search papers, labs, and topics across Lattice.
This paper introduces OpenRCA 2.0, a novel benchmark for root cause analysis (RCA) in large language models (LLMs), addressing the limitations of existing datasets that only label root causes without detailing the causal propagation paths. By employing a step-wise labeling protocol called PAVE, which reconstructs causal pathways through known interventions, the authors provide a more rigorous evaluation framework for LLMs' reasoning capabilities. The findings reveal that while LLMs can identify at least one correct root cause in 76% of cases, they successfully ground this in a verified causal path only 61.5% of the time, highlighting a critical gap in their diagnostic capabilities.
LLMs struggle to connect identified root causes to their causal paths, achieving only 61.5% success in grounding diagnoses despite a 76% identification rate.
Root cause analysis (RCA) poses a holistic test of LLM agentic capabilities, such as long-context understanding, multi-step reasoning, and tool use. However, existing datasets suffer from a fundamental gap: they label only the root cause, not the propagation path connecting it to the observed symptom, which largely simplifies the task to naive pattern matching. To support rigorous evaluation, we introduce PAVE, a step-wise labeling protocol that leverages known interventions from fault injection to reconstruct causal propagation paths. The mechanism is forward verification: reasoning from cause to effect rather than inferring backward from symptoms. Applying PAVE yields OpenRCA 2.0 (500 instances), the first cross-system RCA benchmark with step-wise causal annotations for LLM agents. Across 11 frontier LLMs, recovering the exact root-cause set succeeds in only 20.7% of cases on average. To locate where this difficulty lies, we relax the criterion and find what we call the ungrounded diagnosis: agents identify at least one correct root-cause service in 76.0% of cases, but ground that service in a verified causal propagation path to the observed symptom in only 61.5%. Outcome-only evaluation hides this failure mode; step-wise causal ground truth is the missing piece for trustworthy LLM-based RCA agents.
