Measure the behavior of critical components in a production environment using system metrics that Tamr Core provides to Prometheus and Graphite. This topic is intended for single-node on-premises deployments.
Metrics provided by Tamr Core in Prometheus
Tamr Core provides the following metrics that show changes over time to Prometheus:
- Tamr microservices metrics. Each of the Tamr Core microservices exposes metrics consumable by Prometheus, such as request and error rates; and process statistics, including memory usage, garbage collection (GC), and the number of threads used.
- Elasticsearch cluster metrics. Submitted by the Elasticsearch exporter.
- Node system metrics. Examples include disk I/O statistics, CPU usage, memory usage, and network statistics.
- PostgreSQL database metrics. These metrics include process statistics including memory and CPU usage, and database statistics including buffer statistics, number of connections, cache hit rate, and row insert / fetch / update / delete statistics.
Metrics provided by Tamr Core in Graphite
Tamr Core provides the following Spark metrics to Graphite:
- Application and cluster statistics, including application runtime, the number of active cores and task completion rate.
- Filesystem I/O statistics, memory usage, Codegen compile time, and garbage collection rate.
Updated over 1 year ago