Service usage statistics are collected using a metrics API pod running in the cluster. These metrics are persisted by a metrics API and are exposed through a service metrics viewer pod that queries the metrics API.
Access the metrics viewer app
The metrics viewer app uses Prometheus and Grafana components to create and view custom dashboards consisting of graph panels for various metrics. On the dashboards in the metrics viewer app, you can use PromQL queries on the metrics API to generate graphs for metrics across different periods of time.
To access the metrics viewer app, follow these steps:
- Go to the usagestatistics API in the ArcGIS Enterprise Administrator API directory: https://organization.example.com/<context>/admin/usagestatistics.
- Click the Metrics Viewer resource link.
A new tab opens in the metrics viewer app.
- Sign in to the metrics viewer app.
By default, the initial administrator account credentials are used to access the metrics viewer. You can update these credentials using the Update Credentials operation in admin/usagestatistics. The new password must be a minimum of eight characters and must contain at least one letter (A-Z, a-z), one number (0-9), and one special character.
View statistics dashboards
To view statistics dashboards in the metrics viewer, follow these steps:
- Select Manage from the Dashboard section of the navigation bar to view a list of your dashboards.
ArcGIS Enterprise on Kubernetes includes the dashboards listed in the following section.
- Click the dashboard you want to view.
Included dashboards
The dashboards that are included in ArcGIS Enterprise on Kubernetes are described in the following table:
Caution:
Use save-as to make a duplicate of the dashboard before making any changes to it. Direct changes to default dashboards are not supported
| Dashboard name | Description |
|---|---|
Service usage statistics summary | Provides graph panels for the rate of requests, average response times, failed and successful requests, and usage and wait time-out requests. You can filter by services. |
Service response time percentages | Provides panels that show a percentage of requests with response times lower than a specific service-level agreement (SLA) value. For example, a panel can show the percentage of requests that had a response time of 0.5 seconds or less. You can filter by services and operations. |
Service usage statistics by operation | Provides panels for the rate of requests, average response times, failed and successful requests, and usage and wait time-out requests. You can filter by services and operations. |
Services usage statistics comparison | Provides panels for the rate of requests, average response times, failed and successful requests, and usage and wait time-out requests. You can filter by services and compare them to other services. |
Service usage statistics summary by folder | Provides panels for the rate of requests, average response times, failed and successful requests, and usage and wait time-out requests. You can filter by folders. |
Services usage statistics summary by user | Provides panels for the rate of requests, average response times, failed and successful requests, and usage and wait time-out requests. You can filter by services and users. |
Summary of asynchronous geoprocessing service usage | Provides graph panels for average job submission time in seconds, average job wait time in seconds, rate of job status requests, and average job execution time in seconds. You can filter by asynchronous geoprocessing services |
Dashboard creation
You can also create dashboards using an appropriate PromQL query for specific metrics and filters.
To learn more about creating a dashboard, see the Grafana documentation.