About code scanning
Code scanning es une fonctionnalité que vous utilisez pour analyser le code dans un dépôt GitHub afin de détecter d’éventuelles vulnérabilités de sécurité et erreurs de codage. Tous les problèmes identifiés par l’analyse sont énumérés dans votre référentiel.
You can configure code scanning to run CodeQL analysis and third-party analysis. Code scanning also supports running analysis natively using GitHub Actions or externally using existing CI/CD infrastructure. The bullets below summarize the options available to users when you configure votre instance GitHub Enterprise Server to allow code scanning using actions.
- CodeQL : utilise GitHub Actions avec la configuration par défaut (voir « Définition de la configuration par défaut pour l’analyse du code ») ou la configuration avancée (voir « Configuration de la configuration par défaut pour l’analyse du code »), ou exécute l’analyse CodeQL dans un système d’intégration continue (CI) tiers (voir « Utilisation de l'analyse du code avec votre système CI existant »).
- Third‑party : utilise GitHub Actions ou des outils tiers et charge les résultats dans GitHub Enterprise Server (voir « Chargement d’un fichier SARIF sur GitHub »).
Checking whether your license includes GitHub Advanced Security
Vous pouvez déterminer si votre entreprise dispose d’une licence GitHub Advanced Security en examinant ses paramètres. Pour plus d’informations, consultez « Enabling GitHub Advanced Security for your enterprise ».
Prerequisites for code scanning
-
A license for GitHub Advanced Security (see "À propos de la facturation pour GitHub Advanced Security")
-
Code scanning enabled in the management console (see "Enabling GitHub Advanced Security for your enterprise")
-
A VM or container for code scanning analysis to run in.
Running code scanning using GitHub Actions
Provisioning a self-hosted runner
GitHub Enterprise Server can run code scanning using a GitHub Actions workflow. First, you need to provision one or more self-hosted GitHub Actions runners in your environment. You can provision self-hosted runners at the repository, organization, or enterprise account level. For more information, see "À propos des exécuteurs auto-hébergés" and "Ajout d’exécuteurs auto-hébergés."
If you are provisioning a self-hosted runner for CodeQL analysis, your runner must use a CodeQL-supported operating system version and CPU architecture. For more information, see the CodeQL system requirements.
If you are using default setup for code scanning, assign the code-scanning
label to your self-hosted runner. For more information about using labels with self-hosted runners, see "Utilisation d’étiquettes avec des exécuteurs auto-hébergés." For more information about using default setup for code scanning analysis of compiled languages, see "Analyse du code CodeQL pour les langages compilés."
You must ensure that Git is in the PATH variable on any self-hosted runners you use to run CodeQL actions.
Note: If you use CodeQL code scanning to analyze code written in Python in your enterprise, you must make sure that your self-hosted runner has Python 3 installed.
Provisioning a runner scale set
You can use Actions Runner Controller to create a dedicated runner scale set for your GitHub Enterprise Server instance. For more information, see "Déploiement de groupes identiques d’exécuteurs avec Actions Runner Controller."
Provisioning the actions for code scanning
If you want to use actions to run code scanning on GitHub Enterprise Server, the actions must be available on your appliance.
The CodeQL action is included in your installation of GitHub Enterprise Server. If both GitHub Enterprise Server 3.13 and your GitHub Actions runner have access to the internet, the action will automatically download the CodeQL 2.16.5 bundle required to perform analysis. Alternatively, you can use a synchronization tool to make the latest released version of the CodeQL analysis bundle available locally. For more information, see "Configuring CodeQL analysis on a server without internet access" below.
You can also make third-party actions available to users for code scanning, by setting up GitHub Connect. For more information, see "Configuring code scanning for your appliance" below.
Configuring CodeQL analysis on a server without internet access
If the server on which you are running GitHub Enterprise Server is not connected to the internet, and you want to allow users to enable CodeQL code scanning for their repositories, you must use the CodeQL action sync tool to copy the CodeQL analysis bundle from GitHub.com to your server. The tool, and details of how to use it, are available at https://github.com/github/codeql-action-sync-tool.
If you configure the CodeQL action sync tool, you can use it to sync the latest releases of the CodeQL action and associated CodeQL analysis bundle. These are compatible with GitHub Enterprise Server.
Configuring GitHub Connect to sync GitHub Actions
- If you want to download action workflows on demand from GitHub.com, you need to enable GitHub Connect. For more information, see "Managing GitHub Connect."
- You'll also need to enable GitHub Actions for votre instance GitHub Enterprise Server. For more information, see "Getting started with GitHub Actions for GitHub Enterprise Server."
- The next step is to configure access to actions on GitHub.com using GitHub Connect. For more information, see "Enabling automatic access to GitHub.com actions using GitHub Connect."
- Add a self-hosted runner to your repository, organization, or enterprise account. For more information, see "Ajout d’exécuteurs auto-hébergés."
Running code scanning using the CodeQL CLI
If you don't want to use GitHub Actions, you should run code scanning using the CodeQL CLI.
The CodeQL CLI is a command-line tool that you use to analyze codebases on any machine, including a third-party CI/CD system. For more information, see "Utilisation de l'analyse du code avec votre système CI existant."