Run your app

Deployments and builds

A build turns your repository into an image. A deployment runs that image. Follow both from your service, or use the workspace history to see what changed across your apps.

Workspace history#

Open Deployments in the workspace sidebar. The newest entries come first, including deploys, starts, stops and restarts. Each entry shows the service and project, action or commit, who triggered it, time and status. Finished entries also show their duration.

Use Status and Project to narrow the list. In progress includes Pending, Building and Deploying. Press Clear filters to reset or Load more for older entries. The list refreshes while a loaded entry is in progress. Open an entry to see its deployment progress.

The Builds tab#

Open your service's Builds tab to see its builds and statuses. Open a build for its log, error and deployment progress. Trigger build builds from the connected GitHub repository; a successful build starts a deployment automatically. Developers, Admins and Owners can trigger builds.

Build statusMeaning
QueuedWaiting to build.
FetchingGetting the source from GitHub.
BuildingBuilding the app image or images.
PushingSaving the built images to the registry.
SucceededThe images are ready. Check the deployment to see whether the app is running.
FailedThe build did not finish. Read its log and error before trying again.

New code also builds on a push to your branch. See Automatic deploys.

The Deployments tab#

The service's Deployments tab lists its deploys and service actions. Open an entry for live progress. Redeploy runs the current image again with the latest settings and environment variables. To rebuild from source, use Builds → Trigger build or Settings → Redeploy.

Deployment statusMeaning
PendingQueued for the deployment worker.
BuildingThe deployment is in its build stage.
DeployingThe worker is applying the app or carrying out a service action.
RunningThe deployment, start or restart completed successfully.
StoppedThe stop action completed.
FailedThe operation failed. Read the progress for the cause and any recovery.
CancelledThe deployment was cancelled. Read the progress to see what had already happened.

Developers, Admins and Owners can press Cancel on an in-progress deployment's detail page. It changes to Cancelling… while the worker handles the request. Cancellation may wait for the current step; follow the progress until it finishes.

Automatic recovery after a failed rollout

If a new version fails during rollout, Crowther tries to restore the version that was running before. If there is no previous version, or restoring it also fails, it tries to stop the failed containers. The progress records what happened. A failed deployment does not by itself mean the previous app is offline.

Stop, start and restart#

These controls are on the service's Deployments tab for Developers, Admins and Owners. They open a progress page and appear in deployment history. Wait for an in-progress operation to finish before starting another.

ButtonEffect
StopTakes the running app offline until you start it or deploy it again. Confirm before stopping. It stays stopped even if its server restarts.
StartStarts the stopped containers again, using their existing configuration.
RestartRestarts the running containers as they are. It does not build new code or apply changed environment variables. Expect downtime while the processes restart.

These actions keep container data and named volumes. Stopping is not deleting. For data that must also survive a deploy, use named volumes.