Migrating production infrastructure? Get up to ₹10,000 in migration credits. Apply now
voroaDocs
Operational Controls

Maintenance mode: a holding page instead of a broken site

Turn on maintenance mode for a Voroa service so visitors see a holding page during planned work, use a page of your own, and know what keeps running meanwhile.

What this covers

Maintenance mode serves a holding page to your visitors instead of routing them to your service. Use it for planned work where the application genuinely must not be handling traffic: a database migration, a data move, a cutover. Visitors get one clear message rather than errors, timeouts or half-loaded pages.

It takes effect immediately. Nothing is redeployed, no container is restarted, and your service keeps running exactly as it was, so you can keep working against it while the outside world sees the holding page.

Turning it on

  1. Open the service and go to Settings, then Maintenance Mode.
  2. Switch Maintenance mode on.

The change applies within a few seconds. The switch reflects what is actually in force, so if you come back tomorrow it still shows you what your visitors are seeing.

Turn it off the same way when the work is done.

What visitors get

While maintenance mode is on, every request to your service’s address is answered with the holding page and the status code 503 Service Unavailable. That code is the correct one for planned work: it tells search engines and uptime monitors that the outage is temporary and that they should come back, rather than that the page is gone.

This applies to your Voroa subdomain and to every custom domain on the service.

Using a page of your own

You can serve your own page instead of the Voroa one. In the same section, set Maintenance page to its full address, for example https://status.yourdomain.com/down.

A few things to get right:

  • Host it somewhere other than the service you are maintaining. A page served by the service that is in maintenance cannot load while the maintenance is on. Voroa refuses an address on this service’s own domains for that reason.
  • Make the page self-contained. Voroa fetches the page and serves its HTML. Anything the page loads from a relative path — a stylesheet at /style.css, an image at /logo.png — is resolved against your service’s address, which is answering with the holding page. Inline your styles, or use full https:// addresses for anything the page pulls in.
  • Keep it reachable and small. The page must be on the public internet, served over http or https, respond within five seconds and be under 256 KB.

If your page cannot be fetched for any reason, Voroa serves its own page instead, still with 503. Your visitors always get a holding page, never an error from a fetch that failed.

The fetched page is cached for up to 30 seconds, so an edit on your side takes up to half a minute to appear.

Clearing the field and saving goes back to the Voroa page. Turning maintenance mode off does not clear it — your page is still there the next time you need it.

What maintenance mode does not do

  • It does not let you through. There is no bypass address, and no way to view the real service at its public address while maintenance is on. If you need to verify the work first, do it from a preview environment or against the service’s own address from inside your own tooling before you switch maintenance off.
  • It does not stop your service. The container keeps running and keeps costing what it normally costs. Background work, scheduled tasks and outbound connections all continue. If you want the service actually stopped, suspend it from Settings, Delete or suspend instead.
  • It does not apply to a worker or a cron service. Neither has a public address, so there is nothing to put a holding page in front of, and the setting is not shown for them.
Last updated September 5, 2026