SonarScanner for Maven
The SonarScanner for Maven is a Maven plugin that allows you to execute SonarCloud code analysis via a regular Maven goal.
As a Maven goal, the scanner is available anywhere Maven is available (locally, in CI services, etc.), without the need to manually download, set up, and maintain a separate installation.
Additionally, because the Maven build process already has much of the information needed for SonarCloud to successfully analyze a project, this information is automatically available to the scanner, reducing the amount of manual configuration needed.
Prerequisites
- Maven 3.2.5+
- Java 17 or later
A simple example
In the simplest case, you could perform the analysis manually by invoking the Maven goal, while providing the essential parameters. Something like this:
Usually, you would integrate the mvn
invocation into your build pipeline, to be run on each commit to your repository. The following sections describe how to do this.
Configuration
While the SonarScanner for Maven does automatically detect much of the information needed to perform code analysis, some manual configuration is needed. At the minimum you need to supply the parameters mentioned above: sonar.token
, sonar.host.url
, sonar.organization
, and sonar.projectKey
.
In general, any of these parameters can be configured just like any other maven property (in order of override priority):
- On the
mvn
command line where the scanner goal is invoked, using the-D
argument prefix. - In the
pom.xml
file of the project. Unlike the plain-vanilla SonarScanner CLI, the SonarScanner for Maven uses thepom.xml
instead of thesonar-project.properties
file. - In the global
settings.xml
.
Authentication
sonar.token
: This is a personal access token generated in your SonarCloud account at My Account > Security > Generate Tokens. It allows the scanner to authenticate to SonarCloud. This is usually set via the SONAR_TOKEN
environment variable.
For example, in the GitHub Actions CI environment, you would configure a GitHub Secret called SONAR_TOKEN
with the access token as its value. Then you might have something like the following in your .github/workflows/build.yml
:
The SonarScanner for Maven automatically picks up the value directly from the environment variable. If you use an environment variable, you do not need to pass the token on the mvn
command line.
Server, organization, and project
sonar.host.url
: This is the URL of the SonarCloud server. It is needed because the SonarScanner for Maven plugin also works with the on-premise SonarQube product, where this parameter is set to the URL of the locally installed server. In our case, this parameter should always be set to https://sonarcloud.io
.
sonar.organization
: This is the key of the SonarCloud organization where your project resides. It can be found in the top right of the organization overview page and on the information page of your project.
sonar.projectKey
: This is the key of the SonarCloud project itself, that is, the one resulting from importing the repository that you are now configuring. It can be found on the information page of your project.
These three parameters are usually set on the command line of the mvn
command invoked during your build in your CI environment. For example, in the GitHub Actions CI environment you might have the following in your .github/workflows/build.yml
:
Optional parameters
Additional parameters beyond the required ones can also be set, either
- in the SonarCloud UI,
- in your project
pom.xml
, - or on the command line, as appropriate.
If set on the command line they are simply appended to the mvn
command using additional -D
argument prefixes.
If set in the pom.xml
they are included as part of the project properties. For example:
See Analysis Parameters for an overview of available parameters.
Invoking the goal
When invoking the SonarScanner goal it is recommended that you do it as part of a single maven command in line with the other goals needed for the build. For example:
where the sonar:sonar
goal follows the verify
goal.
This is in contrast to invoking sonar:sonar
in a dedicated mvn
invocation. For example:
The advantage with the first technique is that the SonarScanner has access to the full build context and can therefore make a more thorough analysis. For this reason, the first technique is preferred.
Setting the plugin version
In the pom.xml file
We recommend locking down versions of Maven plugins in the pom.xml
file of the project:
When invoking the goal
When invoking the scanner goal, there are two ways to set the plugin version:
- Using the fully qualified name:
- Using the shorthand
sonar:sonar
instead of the fully qualified name. In this case, the latest plugin version is used:
If your sonar-maven-plugin is not defined by the project’s pom.xml file, we recommend specifying the version instead of using the latest to avoid breaking changes at an unwanted time.
Code coverage
To get coverage information, you will need to generate the coverage report before the analysis and specify the location of the resulting report in an analysis parameter. See Test Coverage for details.
Adjusting the analysis scope
The analysis scope of a project determines the source and test files to be analyzed.
An initial analysis scope is set by default. With the SonarScanner for Maven, the initial analysis scope is:
- For source files: all the files stored under
src/main/java
(in the root or module directories). - For test files: all the files stored under
src/test/java
(in the root or module directories).
To adjust the analysis scope, you can:
- Adjust the initial scope: see below.
- Exclude specific files from the initial scope: see Analysis scope.
- Exclude specific modules from the analysis: see below.
Adjusting the initial scope
The initial scope is set through the sonar.sources
property (for source files) and the sonar.tests
property (for test files). See Analysis parameters for more information.
To adjust the initial scope, you have two options:
- override these properties by setting them explicitly in your build like any other relevant maven property: see Analysis scope.
- use the scanAll option to extend the initial scope to non-JVM-related files. See below.
Using the scanAll option to include non-JVM-related files
You may want to analyze not only the JVM main files but also files related to configuration, infrastructure, etc. An easy way to do that is to enable the scanAll option (By default, this option is disabled.).
If the scanAll option is enabled then the initial analysis scope of source files will be:
- The files stored in
src/main/java.
- The non-JVM-related files stored in the root directory of your project.
The scanAll option is disabled if the sonar.sources
property is overridden.
To enable the scanAll option:
- Set the
sonar.maven.scanAll
property toTrue
. See Analysis parameters.
Excluding a module from the analysis
To exclude a module from the analysis, you may:
- In the
pom.xml
of the module you want to exclude, define the<sonar.skip>true</sonar.skip>
property. - Use build profiles to exclude some modules (like for integration tests).
- Use Advanced Reactor Options (such as
-pl
). For examplemvn sonar:sonar -pl !module2
Troubleshooting
If you get a java.lang.OutOfMemoryError
With SonarScanner for Maven version 5.0 or later
Set the SONAR_SCANNER_JAVA_OPTS
environment variable, like this in Unix environments.
In Windows environments, avoid the double quotes, since they get misinterpreted.
With SonarScanner for Maven version 4.0 or earlier
Set the MAVEN_OPTS
environment variable, like this in Unix environments:
In Windows environments, avoid the double quotes, since they get misinterpreted:
Was this page helpful?