Rule-based incident correlation engine for noisy multi-service alert streams: groups alerts by severity, time window, and service dependency topology, then generates structured root-cause narratives. Validated against ground-truth synthetic incident scenarios (DB cascade, bad deploy, third-party outage).
python observability ai-agents root-cause-analysis alert-triage llm-reasoning sre-tooling incident-correlation service-dependency-graph synthetic-test-data
-
Updated
Jun 20, 2026 - Python