ARISE (Adaptive Retrieval for Incident Support and Explanation): Incident-Triage Datasets

Three datasets for studying memory-augmented incident triage over microservices telemetry and real-world Apache Jira issues, accompanying the paper ARISE: Adaptive Retrieval for Incident Support and Explanation (under double-blind review). online-boutique (6,720 telemetry windows x 347 tickets; full logs/traces/metrics/k8s events on the Google Online Boutique benchmark); otel-demo (1,643 windows x 147 tickets on the OpenTelemetry Demo, a cross-application generalization set); world-of-logs (9,341 real Apache-project Jira tickets, text only). Each archive extracts to a self-describing <dataset-id>/ directory with a dataset card, field schemas, gold relevance relations, and a MANIFEST.sha256.json integrity manifest; SHA256SUMS.txt lists archive checksums. Anonymized for review. The World of Logs view contains only the Apache-Jira subset, derived from The Public Jira Dataset (doi:10.5281/zenodo.15719919, CC BY 4.0); the broader WoL source (The Stack / Stack Overflow / Common Crawl) is not redistributed.