Monitoring a DataStax Enterprise cluster

Information about performance characteristics, thread pool, read/write latency, table statistics, compaction metrics, and endpoint metrics.

DataStax Enterprise (DSE) provides a wealth of metrics regarding your clusters. Understanding the performance characteristics of a cluster is critical to diagnosing issues and planning capacity.

DataStax recommends using DSE Metrics Collector or DataStax Enterprise OpsCenter to monitor clusters and view metrics. These tools provide valuable insights into your clusters by visually representing the most useful metrics in customizable dashboards.

DSE exposes a number of statistics and management operations via Java Management Extensions (JMX). You can get statistics and metrics using DataStax tools or the Java Console (JConsole), and then add those metrics to dashboards in external monitoring tools (such as Prometheus) or OpsCenter.

DSE Metrics Collector and external dashboards

When DSE Metrics Collector is enabled, DSE sends metrics and other structured events to DSE Metrics Collector. Use dsetool insights_config to enable and configure the frequency and type of metrics that are sent to DSE Metrics Collector. After setting the configuration properties, you can export the aggregated metrics to monitoring tools like Prometheus, Graphite, and Splunk, which can then be visualized in a dashboard such as Grafana.

OpsCenter Dashboard and Performance Service

Use OpsCenter to monitor performance metrics in the OpsCenter Dashboard. Real-time and historical performance metrics are available at different granularities: cluster-wide, per node, per table (column family), or per storage tier.

In addition to the OpsCenter Dashboard, you can enable the OpsCenter Performance Service (6.8 | 6.7 | 6.5 | 6.1), which combines OpsCenter metrics with CQL-based diagnostic tables populated by the DSE Performance Service. Use the information generated by the Performance Service to help understand, tune, and optimize cluster performance.