Skip to main content

Deployments And Logs

Application detail pages include Deployments, Live Logs, and Settings tabs.

Deployments Tab

The deployments tab shows every deployment record for the app.

Each row can show:

  • Status.
  • Commit short SHA.
  • Duration.
  • Preview URL.
  • Active marker for static sites.
  • Recovery hint for failures.
  • Logs button.
  • Relative creation time.

Deployment Status

StatusMeaning
pendingDeployment was queued.
buildingMoltern is building the source.
deployingMoltern is releasing the built app.
successThe deployment completed.
failedThe deployment did not complete.

Live Logs

Open Live Logs when:

  • A deploy is still running.
  • A deploy failed.
  • An app crashes after startup.
  • You changed commands or variables.
  • You need to see which build step failed.

Logs are the first source of truth for build and runtime errors.

Recovery Cards

When Moltern can recognize a failure pattern, the detail page may show a recovery card. Use that guidance before retrying.

Common recovery areas:

  • Repository access.
  • Missing variables.
  • Build command failure.
  • Start command failure.
  • Port mismatch.
  • Runtime crash.
  • Quota or plan limit.

Promote A Static Site Deployment

For static site apps, a successful deployment can be promoted to active.

Use promote when:

  • You tested a successful preview and want it live.
  • The latest deployment is not the desired version.
  • You want to restore an earlier static release.

Rollback

Rollback restores a previous active static-site release when available.

Use rollback when:

  • A newly promoted static release has a production issue.
  • The previous release is known to be good.
  • You need a fast recovery path before building a new fix.

For web services, prefer deploying a corrected commit or reverting the source branch and deploying again.

Retry Discipline

Do not retry blindly. Before retrying:

  1. Read the first failing log section.
  2. Identify whether the failure is source, command, variable, port, or quota related.
  3. Fix the cause.
  4. Deploy again.

Repeated deploys without a fix create noise and make incident review harder.