Enabling Code Security for Python

Cette page n'est pas encore disponible en français, sa traduction est en cours.
Si vous avez des questions ou des retours sur notre projet de traduction actuel, n'hésitez pas à nous contacter.

You can detect code-level vulnerabilities and monitor application security in Python applicationss running in Docker, Kubernetes, Amazon ECS, and AWS Fargate.

NOTE: Code-Level Vulnerability detection in Python is in Preview.

Follow these steps to enable Code Security in your service:

  1. Update your Datadog Agent to at least version 7.41.1.

  2. Update your Datadog Tracing Library to at least the minimum version needed to turn on Code Security. For details, see Library Compatibility page.

  3. Add the DD_IAST_ENABLED=true environment variable to your application configuration.

    From the command line:

    DD_IAST_ENABLED=true ddtrace-run python app.py

    Or one of the following methods, depending on where your application runs:

    Update your configuration container for APM by adding the following argument in your docker run command:

    docker run [...] -e DD_IAST_ENABLED=true [...]

    Add the following environment variable value to your container Dockerfile:

    DD_IAST_ENABLED=true

    Update your deployment configuration file for APM and add the IAST environment variable:

    spec:
  template:
    spec:
      containers:
        - name: <CONTAINER_NAME>
          image: <CONTAINER_IMAGE>/<TAG>
          env:
            - name: DD_IAST_ENABLED
              value: "true"

    Update your ECS task definition JSON file, by adding this in the environment section:

    "environment": [
  ...,
  {
    "name": "DD_IAST_ENABLED",
    "value": "true"
  }
]
    1. Restart your service.
    2. To see Code Security in action, browse your service and find code-level vulnerabilities in the Vulnerability Explorer.

    If you need additional assistance, contact Datadog support.

    Third-Party Library Compatibility Note

    Code Security modifies Python code at runtime. This could cause conflicts with other third-party Python libraries that perform similar code transformations, particularly with the following, though not limited to them:

    • Numba
    • JAX
    • TorchScript
    • TensorFlow
    • Bytecode
    • Codetransformer
    • PyPy

    Additionally, Code Security does not correctly propagate taint ranges over native (compiled) code. Therefore, if your codebase heavily relies on modules written in C or C++, using the CPython API, or on intermediate language systems like Cython, the results might be less accurate than expected.

    Further Reading

    Documentation, liens et articles supplémentaires utiles:

    Supported code-level vulnerabilities listDOCUMENTATION more more Enhance application security in production with Datadog Code SecurityBLOG more more Find vulnerabilities in your code with Datadog Code SecurityBLOG more more Datadog Code Security achieves 100 percent accuracy in OWASP Benchmark by using an IAST approachBLOG more more Troubleshooting Application SecurityDOCUMENTATION more more