Metrics & performance
Use telemetry to find slow requests and inspect changes over time.
Choose the right view
| View | When to use it |
|---|---|
| Nodes and telemetry | Check which runtime is reporting and whether its recent activity is healthy. |
| Metrics | Inspect a focused time series or request-performance change. |
| Analytics | Compare broader workspace trends over the available time ranges. |
| Profiles | Inspect runtime activity associated with sampled transactions, when enabled. |
Enable backend performance capture
Errors and transactions have separate sample rates. sampleRate controls errors; tracesSampleRate controls performance transactions. Backend tracing is off by default. Set a transaction sample above zero in Node.js or Go, then verify transactions arrive before enabling profiling.
Investigate a change
Sampling and plan retention affect how much evidence is available. Check the event's release and full commit before associating a performance change with a particular build.
Start with an affected node, environment, and time range.
Look for a latency or volume change around a deployment.
Open the corresponding transaction or timeline and inspect the operations that took time.
Use an available profile for a closer view, then compare the result with your application logs and code.