Documentation

Release flow

Deployments

Deployments move source code into a site release while preserving logs, selected run details, webhook state, and release history.

Updated Jun 22, 2026 2 min read

What deployments do

Deployments move source code into a site release while preserving status, logs, selected run details, webhook state, and release history. A deployment should be treated as an operational record, not just a command that ran once.

Before deploying

  • Confirm the source provider is connected and synced.
  • Review the repository, branch, tag, deploy key, and webhook settings.
  • Check the deployment script for dependency installs, cache commands, migrations, and release activation.
  • Make sure backup and rollback expectations are clear for production sites.

Typical workflow

  1. Connect GitHub or GitLab.
  2. Attach the repository and branch to the site.
  3. Review the deployment script and environment values.
  4. Run a manual deployment or enable push-to-deploy.
  5. Inspect the deployment result, duration, output, and selected run details.

Checks before enabling push-to-deploy

  • The manual deployment path works first.
  • The deploy key can read the repository.
  • Webhooks are attached to the expected branch.
  • The deployment script is safe for production.
  • Notifications are routed to the right team destination.

Common issues

  • A deploy fails because the deploy key is missing or stale.
  • A migration runs without a backup plan.
  • Push-to-deploy is enabled on the wrong branch.
  • Failed deployments are retried without reading the output.

Related guides

Related documentation

Applications

Sites

Sites combine the application, server, source control, domain, SSL, runtime, deployment, and activity history into one workspace.

Read topic

Access

SSH, SFTP, and keys

Separate shell access, file access, provider access, and deployment access so infrastructure remains understandable during handover.

Read topic

Visibility

Activity, realtime, and troubleshooting

Find the resource status, latest run, activity trail, provider metadata, and failure context before changing production infrastructure.

Read topic