The Metrics service is responsible for reading and writing Metric data in Tableau Server. The service is required in order for Metrics to work properly. To make the Metrics service highly available, configure one or more instances of the service to multiple nodes of the cluster. We recommend you configure at least one instance on every node that is running the Application Server (VizPortal).
| Process | |
| Status | Status of the tsm status -v command. For more information, see View Server Process Status |
| Logging | Logs generated by the |
Impact if the Metrics Service fails
The Metrics Service is required for Metrics to function properly on Tableau Server. In order to minimise issues, install multiple instances of the service in your server installation. If you have a single-node instance of Tableau Server, configure at least two instances of the service on your node. If the Metrics Service stops working, any Metrics requests that are in-process will fail. Subsequent requests are routed to a healthy instance of the service.
Metrics Service in a multi-node cluster
The Metrics Service is required for Metrics to function properly on Tableau Server. In order to minimise issues, install multiple instances of the service in your server installation. In multi-node installations, we recommend you configure at least one instance of the Metrics Service on every node that is running the Application Server (VizPortal). This provides redundancy and maximum performance.
Log files for the Metrics service
The Metrics Service creates two sets of log files:
control-metrics*.log. These logs will contain information about the service starting and being enabled.metrics_*.log. Any errors or problems are logged here.
For more information, see Log File Snapshots (Archive Logs).