Analyzing projects for dependencies (SCA)

Advanced Security is only available in SonarQube Server, as an add-on starting in Enterprise Edition.

With Software Composition Analysis (SCA), SonarQube analyzes your built project and returns information on:

• vulnerabilities in your third-party open source dependencies
• where your open source dependencies may conflict with your organization’s license policies

Here are things to check out for to ensure that you get fast and accurate dependency analysis.

Enabling the SCA service

By default, SCA is not enabled on your instance. To turn on the Sonar SCA service as an admin: 

1. Make sure your SonarQube Server license includes Advanced Security.
2. Go to Administration > Configuration > General Settings > Advanced Security and activate the Software Composition analysis (SCA) option.

A connectivity test is available after SCA is enabled to test your Internet access.

Internet connection

Detecting and remediating third-party vulnerabilities requires a constantly updated source of data. New vulnerabilities are discovered every day, and new releases of software that fix them soon follow. Sonar’s researchers are constantly checking to ensure that our license data is accurate, and for details on how reported vulnerabilities may actually affect your code.

As a result, an internet connection is required to always provide the most up-to-date information on your third-party dependencies, including:

• what licenses you have
• what issues you are affected by
• what workarounds maintainers have published as being available

Your SonarQube Server instance must be able to reach the following servers:

•  api.sonarcloud.io
• scanner.sonarcloud.io

You can use the connectivity check under Administration > Configuration > General Settings > Advanced Security to test your network access.

What data is collected

Whenever you run an analysis, data is sent to a Sonar cloud service for analysis. The Sonar scanner collects the manifests of your projects. Manifests are language-specific files that define your projects’ dependencies, such as pom.xml for Java, or requirements.txt for Python. The scanner also collects any relevant lockfiles that describe the fully-resolved set of dependencies, such as package-lock.json for a JavaScript project.

These manifests and lockfiles are assembled into a zip file and sent to a Sonar cloud service for analysis. All data is sent over a secure HTTPS connection. Information on your dependencies and their issues is returned to your SonarQube Server instance. No source code is sent to Sonar. 

Manifests and lockfiles are not stored persistently in Sonar. Sonar may collect aggregate data, and other service telemetry on open source package usage in an anonymized way.

The manifest and lockfiles that are processed contain a list of all dependencies of your project, which could include internally-developed library names. The Sonar service compares dependency names against a set of known open source components; any internally-developed library name would not match, and therefore would not have any license or vulnerability data returned for that library.

Supported languages and package managers

SonarQube evaluates your third-party open source code usage by matching dependencies defined in your project’s dependency files to known open source code on upstream package managers. It currently supports the following languages, package managers, and package manager files: 

Ensure the analysis is run in an appropriate environment

To correctly analyze both your direct and transitive dependencies on projects where there is not a lockfile that contains all dependencies, SonarQube executes commands using your build tools to get a full dependency list.

Note on security

To run a dependency analysis, SonarQube Server might install on its host system all the dependencies required to build your application. This could pull in untrusted artifacts, similar to while you're building the application. Ensure the analysis will run in a secure environment before proceeding. 

Notes on specific build tools and language ecosystems

Maven

The Maven binary (mvn) or maven wrapper (mvnw) must be located in the project directory, the manifest file’s directory, or in the execution path.

Gradle

The Gradle binary (gradle) or Gradle wrapper (gradlew) must be in the project directory, the manifest file’s directory, or in the execution path.

pip 

The analysis must be run with the same Python runtime that your application is built on. The SonarQube analysis will create a virtual environment to resolve dependencies, and a C compiler and development libraries may be required, based on your python dependencies.

Go

The go runtime that matches the version in go.mod must be present.

Internal artifact repositories

If your application build configuration includes internal or private artifacts, the analysis process must have network access to your artifact server.

If the analysis is not run in the proper environment, it will cause degraded analysis results and potential analysis failures. You can see more information in analysis warnings in the UI and in the scanner log. See Troubleshooting for some common scenarios.

Ensure the analysis includes the appropriate files

The SCA analysis recursively searches for appropriate package files for your project. In some cases, this may analyze more files than what your project actually uses. Common cases to look out for include:

• Package manager files in test code and data directories

If you have package manager files present in test directories, ensure these locations are properly excluded from analysis. This can be done in multiple ways:

• Add paths to the common sonar.exclusions configuration option. Example: sonar.exclusions=“tests/**”
• Use the specific sonar.sca.exclusions configuration option. Example sonar.sca.exclusions=“tests/**”
• As long as SonarQubes SCM support is enabled (the default), add the paths to a source control ignore file, such as .gitignore

Customizing the dependency analysis

The following parameters influence the results of the dependency analysis.

Setting up scheduled analyses for SCA

In the initial release of SonarQube Advanced Security, all vulnerabilities and detected licenses are evaluated at analysis time. 

While a future SonarQube release will automatically scan periodically for new vulnerabilities and updated licenses, at the moment a new analysis must be performed to discover newly discovered or newly public vulnerabilities on existing code. 

The easiest way to do this is via the use of scheduled analysis from your DevOps platform. Here is how to set it up for a number of common DevOps platforms.

name: Run SonarQube analysis

on:
  push:
    branches:
      - main
  pull_request:
    types: [opened, synchronize, reopened]

name: Run SonarQube analysis

on:
  push:
    branches:
      - main
  pull_request:
    types: [opened, synchronize, reopened]
  schedule:
cron: ‘23 4 * * *‘ 

Troubleshooting the dependency analysis

See Advanced security troubleshooting for guidance on how to troubleshoot the dependency analysis.

Related pages

• Viewing dependencies
• Reviewing and fixing dependency risks
• Managing license profiles and policies
• Troubleshooting
• Best practices for managing dependency risks