Build a High Performance Remote Storage for Prometheus with Unlimited Time Series
CNCF [Cloud Native Computing Foundation] via YouTube
Overview
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore a high-performance remote storage solution for Prometheus that supports unlimited time series. Learn how to overcome the challenges of high-dimensional and high-cardinality tags in cloud-native metrics. Discover a novel approach that combines encoding target labels into a single column and flattening application labels into independent columns using ClickHouse's columnar storage capabilities. Understand how this method achieves performance comparable to VictoriaMetrics, with up to 10x improvement when querying large time series. Gain insights into a solution that effectively eliminates the concept of time series limitations, enabling cloud-native developers to inject high-cardinality tags into metrics without compromising performance or scalability.
Syllabus
Build a High Performance Remote Storage for Prometheus with Unlimited Time Series - Yang Xiang
Taught by
CNCF [Cloud Native Computing Foundation]