---
title: Export metrics
description: Learn how to export test metrics from Allure Report to InfluxDB and Prometheus for custom graphing, including test status, duration, retries, and categories.
---

# Export metrics (Allure 2)

In a test report directory, the `export` subdirectory contains files that can be imported to InfluxDB or Prometheus for building custom graphs.

## Export to InfluxDB

The `export/influxDbData.txt` file contains aggregate test data in the [line protocol](https://docs.influxdata.com/influxdb/v2/reference/syntax/line-protocol/) format for [InfluxDB](https://www.influxdata.com/). When imported, it created the following measurements.

<table>
<colgroup>
   <col style="width: 25ex;"/>
</colgroup>
<thead>
<tr>
   <th>Measurement</th>
   <th>Tags</th>
</tr>
</thead>
<tbody>
<tr><td>

`launch_status`

</td><td>

One tag per each possible [test status](/docs/test-statuses/): `failed`, `broken`, `passed`, `skipped`, `unknown`.

Each value represents the number of tests with the status.

</td></tr>
<tr><td>

`launch_time`

</td><td>

`duration` — the time it took to run all tests. \
`min_duration` — the duration of the fastest test run. \
`max_duration` — the duration of the slowest test run. \
`sum_duration` — all durations, combined.

The values for `duration` and `sum_duration` may differ significantly if your test framework runs test in parallel or spends significant time preparing for each test.

</td></tr>
<tr><td>

`launch_problems`

</td><td>

One tag per each non-empty [category](/docs/categories/).

Each category's name is converted to lowercase, and the spaces are replaced with underscores, for example: `test_defects`, `product_defects`.

Each value represents the number of tests in the category.

</td></tr>
<tr><td>

`launch_retries`

</td><td>

`retries` — the number of unique test runs. \
`run` — the number of test runs that were [retries](/docs/history-and-retries/#retries) of other test runs.

</td></tr>
</tbody>
</table>

## Export to Prometheus

The `export/prometheusData.txt` file contains aggregate test data in the [text-based format](https://github.com/prometheus/docs/blob/main/docs/instrumenting/exposition_formats.md#text-based-format) format for [Prometheus](https://prometheus.io/). When imported, it created the following metrics.

<table>
<colgroup>
   <col style="width: 25ex;"/>
</colgroup>
<thead>
<tr>
   <th>Prefix</th>
   <th>Metrics</th>
</tr>
</thead>
<tbody>
<tr><td>

`launch_status_*`

</td><td>

One metric per each possible [test status](/docs/test-statuses/): \
`launch_status_failed`, \
`launch_status_broken`, \
`launch_status_passed`, \
`launch_status_skipped`, \
`launch_status_unknown`.

Each value represents the number of tests with the status.

</td></tr>
<tr><td>

`launch_time_*`

</td><td>

`launch_time_duration` — the time it took to run all tests. \
`launch_time_min_duration` — the duration of the fastest test run. \
`launch_time_max_duration` — the duration of the slowest test run. \
`launch_time_sum_duration` — all durations, combined.

The values for `launch_time_duration` and `launch_time_sum_duration` may differ significantly if your test framework runs test in parallel or spends significant time preparing for each test.

</td></tr>
<tr><td>

`launch_problems_*`

</td><td>

One metric per each non-empty [category](/docs/categories/).

Each category's name is converted to lowercase, the spaces are replaced with underscores, and the prefix is added, for example: `launch_problems_test_defects`, `launch_problems_product_defects`.

Each value represents the number of tests in the category.

</td></tr>
<tr><td>

`launch_retries_*`

</td><td>

`launch_retries_retries` — the number of unique test runs. \
`launch_retries_run` — the number of test runs that were [retries](/docs/history-and-retries/#retries) of other test runs.

</td></tr>
</tbody>
</table>