Skip to main content

​
Application metrics

Apps automatically collect and publish the following metrics every 30 seconds:
MetricDescriptionUnit
RequestCountHTTP requests, service calls, page viewsCount
LatencyResponse time per requestMilliseconds
WorkerDemandPending + running workflow jobsCount
WorkerCapacityTotal concurrency across active workersCount
Metrics are dimensioned by:
  • type: HTTP Request, Service Call, Page View
  • statusCategory: Success (2xx–3xx), Warning (4xx), Error (5xx)
This enables drill-down queries such as “error rate for HTTP requests” or “p95 latency for service calls.” Each app publishes to its own isolated metrics namespace — apps cannot access each other’s metrics.

​
Infrastructure metrics

In addition to application-level metrics, the following infrastructure metrics are collected by the cloud provider approximately every 60 seconds:
MetricDescription
CPU UtilizationAverage CPU usage across service instances
Memory UtilizationAverage memory usage across service instances
Running Instance CountNumber of currently running instances
Desired Instance CountNumber of instances requested by the auto-scaler

​
Custom metrics

Apps can be instrumented with custom metrics to track application-specific data (see Custom Metrics guide). Custom metrics are published in the same manner as application metrics — collected every 30 seconds and isolated per app.

​
Database metrics

Database metrics are coming soon. In the interim, BYOI users can monitor database metrics directly through their cloud provider’s native tools.

​
Metrics dashboard

Each app ships with a metrics dashboard accessible at: 
https://{your-app-url}/admin/monitoring/dashboard
Metrics dashboard showing application metrics charts
Access requires the logs:view scope. Assign this scope to admin roles in your app’s access control configuration.

​
Dashboard features

  • Live charts with automatic refresh
  • Time range selection: 5 min, 15 min, 1 hr, 6 hr, 24 hr, or custom range
  • Application tab: Request volume, latency distribution, error rates, worker demand vs. capacity
  • Infrastructure tab: CPU utilization, memory utilization, instance scaling
  • Dimension drill-down: Split any metric by status category, request type, or other dimensions
  • Computed metrics: Error rate percentages, capacity utilization transforms
  • Crosshair synchronization: Hover over one chart to highlight the same timestamp across all charts
  • Overlay metrics: Plot multiple metrics on the same chart for correlation
Application metrics may be delayed up to 1 minute. Infrastructure metrics may be delayed up to 2 minutes.