community/developer-guide/fe-profiler.md

---
title: Generate Flame Graphs with FE Profiler
language: en
description: Use async-profiler to generate flame graphs for the Apache Doris FE and locate performance bottlenecks.
keywords:
    - Apache Doris
    - FE Profiler
    - async-profiler
    - flame graph
    - profile_fe.sh
    - FE performance bottleneck
---

<!--
Licensed to the Apache Software Foundation (ASF) under one
or more contributor license agreements.  See the NOTICE file
distributed with this work for additional information
regarding copyright ownership.  The ASF licenses this file
to you under the Apache License, Version 2.0 (the
"License"); you may not use this file except in compliance
with the License.  You may obtain a copy of the License at

  http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing,
software distributed under the License is distributed on an
"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
KIND, either express or implied.  See the License for the
specific language governing permissions and limitations
under the License.
-->

# Generate Flame Graphs with FE Profiler

<!-- Knowledge type: Tool usage -->
<!-- Applicable scenarios: Performance tuning / Troubleshooting -->

## Introduction

In Apache Doris 2.1.4 and later, the FE deployment directory `${DORIS_FE_HOME}/bin` ships with a `profile_fe.sh` script. The script uses [async-profiler](https://github.com/async-profiler/async-profiler) to generate a flame graph of the FE so you can locate FE performance bottlenecks.

### System Support

| System | Supported |
|--------|-----------|
| Linux | Yes |
| macOS | Yes |
| Windows | Not yet |

## Usage Steps

### 1. Run profile_fe.sh

```shell
# Profile the FE for 10 seconds by default and generate a flame graph
${DORIS_FE_HOME}/bin/profile_fe.sh

# Profile the FE for 30 seconds and generate a flame graph
PROFILE_SECONDS=30 ${DORIS_FE_HOME}/bin/profile_fe.sh
```

| Environment Variable | Meaning | Default |
|----------------------|---------|---------|
| `PROFILE_SECONDS` | async-profiler sampling duration (seconds) | `10` |

### 2. Trigger the Workload

After you run the command, async-profiler keeps sampling for the specified duration. During this period, send the target queries repeatedly to the FE so that async-profiler can capture stack frames along the target code paths.

### 3. View the Flame Graph

When sampling finishes, the script generates an HTML flame graph file in the `${DORIS_FE_HOME}/log` directory. FE performance bottlenecks usually appear as "plateaus" (short and wide). Optimize the code corresponding to those stack frames.

Example flame graph:

![](/images/fe-profiler.png)

## Using async-profiler on Older FE Versions

For Doris versions earlier than 2.1.4, download the following two files manually:

1. Download [ap-loader-all-3.0-8.jar](https://repo1.maven.org/maven2/me/bechberger/ap-loader-all/3.0-8/ap-loader-all-3.0-8.jar) into the `${DORIS_FE_HOME}/lib` directory.
2. Download [profile_fe.sh](https://raw.githubusercontent.com/apache/doris/master/bin/profile_fe.sh) into the `${DORIS_FE_HOME}/bin` directory.
3. Run `profile_fe.sh` to generate the flame graph.

## FAQ

**Q: The flame graph does not contain stack frames for the target function during sampling.**

The corresponding query path was usually not triggered during sampling. While `profile_fe.sh` is running, send the target SQL repeatedly by hand or with a script, or extend the sampling duration with `PROFILE_SECONDS=30`.

**Q: How do you choose between FE Profiler and [Arthas](./arthas)?**

- `profile_fe.sh`: a script-based approach with a fixed sampling duration, suitable for quickly generating a one-off flame graph.
- Arthas: an interactive terminal that supports `trace`, `watch`, and other finer-grained runtime observations, suitable for deeper investigation.