Explore efficient methods for collecting Cassandra performance metrics in this 28-minute conference talk by Hayato Shimizu from AxonOps. Learn about the challenges of scraping thousands of metrics per node through the JMX interface, and how this can impact CPU performance. Discover the right approaches used by various metrics collection tools, as well as an innovative "dirty" method that significantly improves performance visibility for Cassandra operators. Gain insights into overcoming constraints on scraping frequency and achieving high-resolution insights into Cassandra cluster performance, especially crucial for mission-critical applications.