Auto-Push for Git Integration

Introduction

Auto-push automatically pushes data contract changes to Git whenever a data contract is saved — either through the UI or the API. This removes the need to manually push each change and keeps your Git repository continuously in sync with Entropy Data.

Auto-push builds on the existing Git Integration and supports all configured Git connection types (GitHub, GitLab, Azure DevOps, Bitbucket Cloud).

Auto-push is disabled by default and must be activated by setting the environment variable APPLICATION_AUTOPUSH_ENABLED to true. Once activated, the auto-push options become available in the organization Git settings and on individual Git connections.

Auto-Push Modes

Configuring Auto-Push

Organization-Level Setting

The organization-level setting defines the default auto-push behavior for all Git connections in your organization.

1. Navigate to Organization Settings → Git Connections
2. In the Git Settings card, select the desired Auto Push mode
3. Click Save

This setting applies to all Git connections that use the default auto-push mode.

Per-Connection Override

Each Git connection can override the organization-level auto-push setting. This allows you to enable auto-push for specific data contracts while keeping it disabled for others.

1. Open a Git-connected data contract
2. In the Git connection card in the sidebar, click Edit
3. Select the desired Auto Push mode:
   • Default (use organization setting) — Inherits the organization-level setting
   • Disabled — Explicitly disables auto-push for this connection
   • Auto push to MR/PR — Explicitly enables auto-push for this connection
4. Click Save

When auto-push is active, the Git connection card displays a badge indicating the current mode.

How It Works

When a data contract is saved and auto-push is enabled:

1. The system resolves the effective auto-push mode by checking the per-connection setting first, then falling back to the organization-level setting.
2. Git credentials are resolved using the standard credential resolution hierarchy.
3. The data contract YAML is pushed to Git.
4. Commit messages, branch names, and PR/MR titles use the templates configured in your Git Settings.

Auto Push to MR/PR

In this mode, each save creates or updates a pull request (or merge request on GitLab):

• If an open PR/MR already exists for the connection, the existing PR/MR branch is updated with the new content.
• If no open PR/MR exists, a new branch and PR/MR are created.

Conflict Detection

Auto-push includes built-in conflict detection to prevent overwriting changes made directly in Git. Before pushing, the system compares the last known state with the current remote state:

• Remote is ahead: If the remote file has been modified since the last sync, the auto-push is skipped to avoid overwriting those changes. You can resolve the conflict manually and then push.
• Already up to date: If the content in Git already matches the data contract, the push is skipped to avoid unnecessary commits.

This ensures that auto-push never silently overwrites work done by other contributors directly in the repository.

API Support

Auto-push also works when data contracts are saved via the API. The commit message will include the API key's display name as the author. Credential resolution follows the same rules — an API key scoped to a user resolves credentials starting at the user level, a team-scoped key starts at the team level, and an organization-scoped key uses organization credentials.

Important Notes

• If a push fails (e.g., due to credential issues or network errors), the data contract save still succeeds. The next sync status check will reflect the current state.
• Auto-push respects the same branch and repository settings configured on the Git connection.