> ## Documentation Index
> Fetch the complete documentation index at: https://www.aptible.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Metrics

> Learn about container metrics on Aptible

## Overview

<img src="https://mintcdn.com/aptible/gJr2xlqbHzeeHUse/images/Metrics-overview.png?fit=max&auto=format&n=gJr2xlqbHzeeHUse&q=85&s=452990249d4fdd19b0dd1c4a0c8d83fa" alt="" width="10368" height="5184" data-path="images/Metrics-overview.png" />

Aptible provides key metrics for your app and database containers, such as memory, CPU, and disk usage, and provides them in two forms:

* **In-app metrics:** Metric visualizations within the Aptible Dashboard, enabling real-time monitoring

* **Metric Drains:** Send to a destination of your choice for monitoring, alerting, and long-term retention

Aptible provides in-app metrics conveniently within the Aptible Dashboard. This feature offers real-time monitoring with visualizations for quick insights.

The following metrics are available within the Aptible Dashboard

* Apps/Services:

  * Memory Usage

  * CPU Usage

  * Load Average

* Databases:

  * Memory Usage

  * CPU Usage

  * Load Average

  * Disk IO

  * Disk Usage

### Accessing in-app metrics

To access metrics in the Aptible Dashboard, select your app or database and click the **Metrics** tab.

<Info>
  Metrics are only available for apps and databases with at least one container running. If a service is scaled to zero containers, metrics will not be displayed.
</Info>

### Interpreting CPU Metrics

The Aptible Dashboard shows raw CPU usage. This value can exceed `100%` because it represents actual CPU consumption, not a normalized percentage of the Container's CPU limit.

Metric drains can report the same underlying CPU data differently depending on the destination. For example, in Datadog you may want to normalize CPU usage against the Container limit when building dashboards or alerts:

```text theme={null}
cpu% = enclave.milli_cpu_usage / enclave.milli_cpu_limit * 100
```

Normalized CPU values are useful for alerting because they remain comparable when you scale services. Raw CPU values are useful when you need to understand actual CPU consumption.

## Metric Drains

Metric Drains provide a powerful option for routing your metrics data to a destination of your choice for comprehensive monitoring, alerting, and long-term data retention.

<CardGroup cols={2}>
  <Card title="Datadog" icon="book" iconType="duotone" href="https://www.aptible.com/docs/datadog" />

  <Card title="InfluxDB" icon="book" iconType="duotone" href="https://www.aptible.com/docs/influxdb-metric-drain" />

  <Card title="Collect Prometheus Metrics" icon="book" iconType="duotone" href="/how-to-guides/observability-guides/collect-prometheus-metrics-from-apps" />
</CardGroup>
