Best practices: Use source control

  • Freigeben Version: Australia
  • Aktualisiert 12. März 2026
  • 2 Minuten Lesedauer
  • Source control (Git) combined with the Application Repository is the preferred deployment method for custom scoped applications. System Update Sets remain functional, but they are the least recommended deployment mechanism for application development.

    Why use source control

    Full change history
    Git provides a complete, auditable record of every change, such as who made it, when, and why, with the ability to use diff, review, and revert at the individual file level.
    Branching and parallel development
    Multiple developers can work on different features simultaneously without conflicting update sets.
    Code review workflows
    Pull requests and merge reviews enforce peer review before any change is published, reducing errors and security oversights.
    Versioned releases
    Tags and branches map directly to Application Repository versions, enabling reliable rollbacks and hotfix workflows.
    CI/CD integration
    External CI/CD tools such as Jenkins, GitHub Actions, Azure DevOps can trigger builds, tests, and deployments using the ServiceNow SDK and CLI.

    When are System Update Sets appropriate

    System Update Sets retain value for quick operational changes to the following:
    • Global-scope configurations.
    • Emergency hotfixes when the full source control pipeline would introduce unacceptable delay.
    • Organizations still transitioning from legacy development practices.
    • Changes to ServiceNow® Store or plugin applications where the Application Repository workflow doesn’t apply natively.
    Even in these scenarios, consider using Instance Scan checks on your System Update Sets before committing them, and batch related update sets to reduce deployment complexity.

    Security best practices

    Enforce least privilege through Access Control Lists
    Define custom application-specific roles rather than relying on broad system roles. Configure Access Control Lists to use the evaluation order of roles first, then conditions, then scripts for optimal performance and security.
    Use deny-by-default
    The glide.sm.default_mode should remain in Deny mode so that wildcard Access Control Lists restrict access to admin-only by default.
    Run Instance Scan automatically
    Configure Instance Scan to run against System Update Sets and application deployments. This catches empty Access Control Lists, overly permissive access, and coding standard violations before they reach production.
    Protect cross-scope access
    Configure application cross-scope access privileges deliberately. Do not default to open access for all scopes.
    Implement Multi-Factor Authentication and IP restrictions
    All interactive accounts performing deployments should use multifactor authentication. Service accounts for automated deployments should be restricted by IP range and have tightly scoped roles.
    Audit everything
    Use Access Analyzer to regularly review user, group, and role permissions. Forward deployment logs and Access Control activity to your SIEM platform for monitoring.
    Separate data from code
    Application data records are not included in Application Repository deployments. Manage reference data and seed data through separate, controlled processes with appropriate data classification.

    Deployment method decision matrix

    The following table shows the scenario and recommended deployment approach:

    Tabelle : 1. Deployment method approach
    Scenario Recommended approach
    New custom scoped application: Use source control + Application Repository + App Engine Management Center Pipelines and Deployments, or ReleaseOps.
    Global scope configuration changes: Use System Update Sets with Instance Scan or global application using the Application Repository.
    Emergency production hotfix: Use System Update Sets for speed and back port into source control (Git) immediately after.
    Citizen developer app from App Engine Studio: Use App Engine Management Center Pipelines and Deployments with a guided approval workflow.
    Multi-team release coordination: Use ReleaseOps with release trains and playbook validation.