Synchronize configuration

Once Source Control is enabled, Casewhere keeps the Git-tracked folder and the running configuration in sync. This article explains the three ways configuration moves between them and the everyday Git workflow that ties environments together.

Automatic export on save

Every change you make in Casewhere Admin is written to the folder automatically as part of the same save. When you create, edit, or delete a resource, its JSON file is created, updated, or removed accordingly — so the folder always reflects the current configuration without any extra step.

Export runs after the change has been committed to the runtime, and a failure to write the folder is logged but never blocks your save. If the folder and runtime ever drift (for example after a crash, or a bulk migration), use Export all to folder to reconcile them.

The Source Control page

Open Deployment → Source Control. Two toolbar actions operate on the whole folder:

Action Direction What it does When to use
Export all to folder Casewhere → folder Writes the current configuration of every product and resource into the folder. Initial seed, or to reconcile after drift.
Reload from folder Folder → Casewhere Imports resources whose folder content differs from the running configuration. The folder wins. After an external git pull brought in changes.

The Last result area reports the outcome — the totals processed, how many were imported or exported, and any errors.

Reload from folder applies changes to live configuration, and the folder version wins on any conflict. Only resources whose content actually differs are re-imported (audit metadata is ignored in the comparison), so a reload is safe to run repeatedly, but always reload from a known, reviewed commit.

Synchronize one resource: Fetch From Folder

You don't always need a full reload. On any management grid (Data Classes, Workflow Definitions, Rules, and so on), select a resource and choose DevOps → Fetch From Folder to reload just that one resource from the folder. Casewhere locates the file by its id and re-imports it only if its content differs from the running configuration. This is handy when you have pulled a change to a single resource and want to bring just that one into the runtime.

The everyday Git workflow

Git operations themselves are performed with your own tooling. A typical multi-environment loop:

sequenceDiagram
    participant AdminA as Admin (environment A)
    participant FolderA as Folder (environment A)
    participant Remote as Git remote
    participant FolderB as Folder (environment B)
    participant AdminB as Admin (environment B)
    AdminA->>FolderA: Save change in Admin (auto-exported)
    FolderA->>Remote: git commit & push (external)
    Remote->>FolderB: git pull (external)
    FolderB->>AdminB: Source Control → Reload from folder
  1. Make changes in Casewhere Admin on environment A. They are exported to the folder automatically.
  2. Commit and push the folder with your Git tooling (or let CI do it). Review changes as a pull request if you wish — every resource is a readable JSON file.
  3. On environment B, pull the changes with Git.
  4. In environment B's Admin, run Reload from folder (or Fetch From Folder for a single resource) to apply the pulled configuration to the runtime.

Best practices

  • Reload from a reviewed commit. Because the folder wins, only reload after you have pulled a known-good revision.
  • Reconcile, don't hand-sync. If you are unsure whether the folder matches the runtime, run Export all to folder to bring the folder up to date, then commit.
  • Keep secrets out of the repository. Configuration JSON can reference environment-specific values — manage environment differences through your deployment settings, not by committing per-environment secrets.
  • Pair with DevOps packages for releases. Use Source Control for history and review, and DevOps packages to promote a curated release between environments.
  • Combine with AI Integration. When you build configuration with AI Integration, Source Control captures exactly what changed so you can review and roll back AI-assisted edits like any other commit.