# Monitoring instances


A monitoring instance can be installed to collect and analyze metric data generated by the [metrics plugin](/monitoring/metrics/index.md) installed in NServiceBus endpoints. This data is exposed to [ServicePulse](/servicepulse/index.md) via an HTTP API.

> [!NOTE]
> The ServiceControl HTTP API is designed for use by ServicePulse only and may change at any time. Use of this HTTP API for other purposes is discouraged.

```mermaid
graph LR
  subgraph Endpoints
    Audit
    Error
    Plugin[Metrics<br>Plugin]
  end

  Plugin -- Metric<br>Data --> IQ

  IQ[Monitoring<br>Input Queue] --> MI[Monitoring<br>Instance]
  ServicePulse -.-> MI
```

> [!NOTE]
> Monitoring instances store data about each endpoint in memory for 10 minutes. Monitoring instances do not store data in a persistent store.
> Restarting the monitoring instance will clear the in-memory cache.

Each environment should have a single monitoring instance that all endpoints are configured to use.