ORCA uses observability data to auto-repair microservice failures
ORCA transforms telemetry into automated code patches—no manual debugging needed
Researchers present ORCA, an observability-grounded automated program repair pipeline for microservice incidents. It distills paired failure and reference telemetry into a fault signature, identifies candidate code and deployment-configuration locations, and uses repair graph agents plus an Exploration agent to generate unified-diff patches. A Telemetry-Grounded Patch Verifier evaluates patches by separating patch validity, syntactic and semantic correctness, test-oracle integrity, and telemetry replay. On a 575-case benchmark, ORCA outperforms all evaluated baselines in cost-effectiveness, showing operational telemetry can be transformed from diagnostic evidence into actionable repair context.
- ORCA (by HKUST researchers) uses telemetry data to auto-generate 575 tested patches for microservice failures
- Repair graph agents + Telemetry-Grounded Patch Verifier validate patches for syntax, semantics, and telemetry replay
- Outperforms all baselines in cost-effectiveness on real-world incident benchmarks
Why It Matters
Replaces manual debugging with AI-driven repair, cutting incident resolution time from hours to minutes for SRE teams.