ops-lite
Summary
ops-lite is a compact root-cause-analysis evaluation set with per-case causal-graph ground truth for microservice systems.
Official Artifacts
- Official Hugging Face dataset: https://huggingface.co/datasets/anon-ops/ops-lite
- Dataset metadata snapshot: ops-lite-2026
Dataset Shape
ops-lite has 500 cases across Train-Ticket, Hotel Reservation / DeathStarBench, and OpenTelemetry Demo. Each case contains normal and abnormal metric windows, injection metadata, environment snapshots, labels, and a causal_graph.json file.
Role In The Wiki
ops-lite is useful when the evaluation target is causal graph/path scoring for RCA, especially in a compact public artifact. It is less suited for long continuous forecasting pretraining.
Relation To Foundation TSFM Agenda
Use the source-level agenda mapping in ops-lite-2026 rather than duplicating verdict rows here.
At the entity level, ops-lite is useful when the evaluation target is causal graph/path scoring for RCA, especially in a compact public artifact. It is less suited for long continuous forecasting pretraining. This page should stay as the object card; source pages carry slot-level verdicts, evidence, and missing pieces.