Atlassian Automates Root Cause Analysis by Correlating Metrics, Logs and Traces

Atlassian introduced an approach to automate root cause analysis in cloud systems by correlating metrics, logs, distributed traces, and service topology to generate ranked hypotheses about the cause and propagation path of failures. The work was published via CNCF.

• System builds service dependency map from real traffic via OpenTelemetry
• Anomalies in metrics, traces, and logs are unified and grouped by time
• Output: ranked hypotheses with probable failure source and evidence
• Atlassian explores LLM orchestration for iterative incident investigations