Skip to content

Maintenance mode

App instance maintenance mode temporarily replaces enabled public HTTP and HTTPS routes with a Wodby maintenance response. Use it when visitors should see a consistent maintenance page while you continue running deployments, migrations, or other work inside the application.

Maintenance mode changes routing only. It does not stop application workloads or change the app instance lifecycle status.

Enable maintenance mode

  1. Open Apps > [App] > [Instance] > Settings > Maintenance.
  2. Select Enable maintenance mode.
  3. Review the affected traffic and confirm.
  4. Follow the routing task if the dashboard opens it.

The maintenance status moves from Off to Enabling while Wodby applies the routing change, then to Active after the public edge confirms it. The app instance keeps its normal lifecycle status, such as running, deploying, awaiting, or errored. Instance headers and lists show a separate maintenance badge.

Enabling maintenance mode requires a non-infrastructure app instance with ready platform-managed routing. You can change maintenance mode while the instance is running, awaiting a deployment, errored, or deploying. The cluster must support the direct-response routing capability used for the fixed page. If the capability is unavailable, the dashboard leaves maintenance mode off and explains the requirement.

Traffic behavior

While maintenance mode is active, every enabled public Wodby HTTP or HTTPS route for the app instance returns the same fixed page with:

  • HTTP status 503 Service Unavailable
  • Retry-After: 300, asking clients to retry after five minutes
  • Cache-Control: no-store
  • search-engine no-index headers

The page is served at the public edge and does not depend on an application container. It is fixed and cannot be customized. Existing hostname and path matching, TLS certificates, HTTP-to-HTTPS redirects, and HTTP basic authentication remain in place.

Maintenance mode also covers public redirect routes. A route configured to redirect to another destination returns the maintenance response until maintenance mode is disabled.

The following traffic is not replaced:

  • routes marked private
  • routes suppressed because App Access publishes them through a protected or private provider path
  • published TCP or UDP ports

With Selected endpoints App Access, any ordinary public routes outside the selected scope still receive the maintenance response. Provider-managed protected or private endpoints continue using their normal provider path.

Workloads, tasks, and billing

Application workloads remain running. Scheduled jobs, builds, deploys, backups, and other application operations continue according to their normal rules. Maintenance mode does not reduce app-service plan usage or infrastructure costs.

A workload deployment and a maintenance request can overlap. When necessary, Wodby waits for the workload deployment before applying the pending routing change. Follow the routing task to see its progress or failure details.

You cannot pause an app instance while maintenance mode is requested or active. Disable maintenance mode and wait for its status to return to Off, then use Settings > Pause & Resume. See Pausing and resuming an instance.

Disable or reverse maintenance mode

Open Settings > Maintenance and select Disable maintenance mode. The status changes to Disabling while Wodby restores application backends to the public routes, then returns to Off after the public edge confirms the change. Visitors may continue to receive the maintenance response during this transition.

You can reverse a pending request. For example, select Disable maintenance mode while the status is Enabling, or select Enable maintenance mode while it is Disabling. Wodby finishes or supersedes the in-progress routing change and converges on the latest requested state.

If a routing task fails, the requested and active states can differ. The dashboard continues to show Enabling or Disabling instead of reporting a completed change. Review the task logs, correct the routing or cluster problem, and retry the failed routing task.

Maintenance mode compared with pausing

Behavior Maintenance mode Paused instance
Public HTTP and HTTPS routes Fixed 503 maintenance response Do not respond
Application workloads Continue running Stop
Scheduled jobs Continue normally Stop and skip runs that become due
Builds and deploys Continue normally Cannot start
Persistent data Remains provisioned Remains provisioned
App-service plan usage Unchanged Stops after pause completes
TCP and UDP ports Unchanged Do not respond with the stopped workloads

Use maintenance mode when the application must keep running behind a temporary public response. Use pause when the instance is not needed and its workloads should release compute and memory.