All Products
Search

This Product

    Function Compute:Add the ARMS Python agent extension to a custom runtime

    Document Center

    Function Compute:Add the ARMS Python agent extension to a custom runtime

    Last Updated:Sep 29, 2025

    This topic describes how to manually install the ARMS agent for a Python function that uses a custom runtime.

    Background information

    Function Compute seamlessly integrates with Application Real-Time Monitoring Service (ARMS). After you install the agent for a Python function, ARMS starts to monitor the Python application. You can use the ARMS platform to view monitoring data such as application topology, call chains, and SQL analysis. For more information, see What is Application Real-Time Monitoring Service (ARMS)?.

    Limits

    The Python version of the function must be the same as the Python version of the custom runtime.

    Step 1: Create an agent layer

    1. Obtain the Python agent package

    1.1. Download the agent package

    // Download a specific version of the agent over a VPC network
wget http://arms-apm-${regionID}.oss-${regionID}-internal.aliyuncs.com/aliyun-python-agent/${version}/aliyun-python-agent.tar.gz
// Download the latest version of the agent over a VPC network
wget http://arms-apm-${regionID}.oss-${regionID}-internal.aliyuncs.com/aliyun-python-agent/aliyun-python-agent.tar.gz
// Download a specific version of the agent over the internet
wget http://arms-apm-${regionID}.oss-${regionID}.aliyuncs.com/aliyun-python-agent/${version}/aliyun-python-agent.tar.gz
// Download the latest version of the agent over the internet
wget http://arms-apm-${regionID}.oss-${regionID}.aliyuncs.com/aliyun-python-agent/aliyun-python-agent.tar.gz

    1. Replace ${regionID} with the region ID. For more information, see Supported regions.

    2. Replace ${version} with the version number. For more information, see Python Agent version guide.

    The following example shows how to download a specific version of the agent:

    wget http://arms-apm-cn-hangzhou.oss-cn-hangzhou.aliyuncs.com/aliyun-python-agent/1.5.0/aliyun-python-agent.tar.gz

    1.2. Decompress the agent package and run pip install

    ## Decompress the package.
tar -zxvf aliyun-python-agent.tar.gz

## Build the package.
pip3 install target/*.whl -t aliyun-instrument

## Modify the aliyun-instrument file in the ./aliyun-instrument/bin folder. Replace the first line #!/usr/bin/python3 with the following content.
#!/var/fc/lang/python3.10/bin/python3

    2. Create a layer in the console

    1. Log on to the Function Compute console. In the left-side navigation pane, choose Advanced Features > Layers.

    2. In the top navigation bar, select a region. On the Layers page, click Create Layer.

    3. On the Create Layer page, select Upload From Folder to upload the layer.

      image

    Note

    When you upload the aliyun-instrument file, make sure that aliyun-instrument is the top-level folder. Do not nest the aliyun-instrument folder within another folder of the same name during compression and decompression. Otherwise, the environment variable configuration is affected.

    Step 2: Use the custom layer in a function

    1. Add the custom layer in the advanced configuration of the function

    1. Log on to the Function Compute console. In the left-side navigation pane, click Functions.

    2. In the top navigation bar, select a region. On the Functions page, click the function that you want to manage.

    3. On the function details page, click the Configuration tab. Click Edit to the right of Advanced Configuration. In the Advanced Configuration panel, choose +Add Layer > Add Custom Layer. Select the Custom Layer and Layer Version. Then, click Deploy.

    image

    2. Add environment variables

    Variable

    Value

    ARMS_APP_NAME

    FC:{function_name}

    ARMS_LICENSE_KEY

    Obtain the LicenseKey

    ARMS_REGION_ID

    {region}

    LD_LIBRARY_PATH

    /code:/code/lib:/usr/local/lib:/opt/lib:/opt/php8.1/lib:/opt/php8.0/lib:/opt/php7.2/lib

    PATH

    /code/python/bin:/var/fc/lang/python3.10/bin:/usr/local/bin/apache-maven/bin:/usr/local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/local/ruby/bin:/opt/bin:/code:/code/bin

    PYTHONPATH

    /opt:/opt/python:/code

    PYTHONUSERBASE

    /code/python

    3. Modify the startup command

    On the function details page, click the Configuration tab. Click Edit to the right of Basic Configuration. Modify the value of Startup Command as follows.

    aliyun-instrument python3 app.py

    If the application is started using uvicorn, replace the startup command with the following command to connect to the agent.

    For example:

    uvicorn -w 4 -b 0.0.0.0:8000 app:app

    Change it to:

    aliyun-instrument gunicorn -w 4 -k uvicorn.workers.UvicornWorker -b 0.0.0.0:8000 app:app

    The aliyun-instrument command handles the initialization and automatic instrumentation of the ARMS Python agent. If your program uses gevent coroutines, you must set the GEVENT_ENABLE environment variable to true. For example, if your program contains the following code: 

    from gevent import monkey
monkey.patch_all()

    Set the environment variable as follows:

    GEVENT_ENABLE=true

    Verify the results

    After you execute the function, wait for about one minute. If the Python application appears on the Application Monitoring > Application List page in the ARMS console and reports data, this indicates that the agent is connected.

    image