1.9 KiB
1.9 KiB
tags:: readingnotes title:: @A Metric Recommendation Service for Online Systems using Graph Learning pages:: 12 item-type:: journalArticle original-title:: A Metric Recommendation Service for Online Systems using Graph Learning language:: en authors:: Anonymous Author library-catalog:: Zotero links:: Local library, Web library
- Abstract
- Today’s monitoring and failure management mechanisms for online systems heavily rely on metrics, which are time series data that can describe the real-time state of a system from various perspectives. Though several attempts have been devoted to automatic failure management based on metrics, the primary step, metric selection, remains manual to a large extent. To better understand the prior practice, we conduct an empirical study on the selected metrics in prior work and obtain some findings. Based on the findings, we develop a metric recommendation service for online systems, which can automate the metrics selection practice and greatly ease the burden in managing an online system. Specifically, we analyze the needs of two key failure management tasks, i.e., anomaly detection and fault diagnosis, and design metric recommendation mechanisms for them respectively. Graph learning techniques are employed in the automation of metric recommendation. Our experiments demonstrate that the proposed approach can achieve an F1score of 0.912 in selecting metrics for anomaly detection, and an accuracy of 0.859 in retrieving metrics for faults diagnosis, which significantly outperforms the compared baselines.
- Attachments
- Author - A Metric Recommendation Service for Online Systems.pdf {{zotero-imported-file CVQ64RBK, "Author - A Metric Recommendation Service for Online Systems.pdf"}}