Cancelling hosting
Cancelling is not the same as erasing. You agree a date on which service ends, your data stays recoverable for the retention you already accepted, and the deletion itself is yours to confirm — nobody here approves it for you.
You will find Cancelling hosting under Hosting → Settings → Cancellation, as two numbered steps: check what cancelling stops, then request it.
Read the review first
The review is the whole point of this screen. Before anything is requested, it states:
- the applications and services that stop serving;
- the databases and files affected;
- the service-end date you choose;
- the date your data stops being recoverable;
- the route to get your data out;
- anything stopping the deletion, named with the lock that is holding it.
The reviewed scope is stamped, so what you confirm is exactly what was shown to you. If the scope changes afterwards, the confirmation no longer describes it and the platform will not delete until you review and confirm the new one.
Choose what happens at the end
When you request cancellation you choose one of two things:
- Delete at the service-end date. Service keeps running until that date, then the environment is deleted. This is the ordinary choice.
- Delete now. Service ends immediately and the environment is deleted as soon as the change is applied. You can also choose this on a cancellation you already made, if you decide you do not want to wait.
Both choices show the same review, and both require you to confirm that you have read it. Neither waits for a person on our side.
Request cancellation
- Open Settings → Cancellation.
- Read the review.
- Give a reason.
- Choose a service-end date inside the window the form offers.
- Choose whether to delete at that date or now.
- Confirm that you have reviewed the scope, and submit.
The request can be withdrawn until deletion starts. Withdrawing restores the desired service state; access returns after the change is applied. Once deletion has begun it cannot be withdrawn.
What happens at service end
On the agreed service-end date:
- the applications stop serving — their running copies go to zero and their routes are withdrawn;
- then the environment is deleted: its namespace and application resources, its storage, its addresses, the names answering for it, its private build objects and its site identity are all removed, in that order, and each step is recorded on the cancellation.
No step waits for us. The platform does this on its own schedule.
What is kept, and for how long
Deletion removes the environment, not your backups. Both backup tracks — database and files — are kept for the retention agreement in force, until the recoverable-until date the review showed. After that date they are pruned on their own schedule and the data is gone.
One thing worth being precise about: that date is the retention term on record. The review also tells you whether we have actually confirmed a restore point exists for the environment. If it says we have not been able to confirm one, treat the date as the deadline rather than as a copy that is waiting for you, and ask for an export before you confirm.
Delete everything now, instead
Retention is the default, not a waiting period you are stuck with. If you want the data gone rather than recoverable for a week, Delete everything now destroys the environment and its retained backups at the moment you confirm.
That is erasure, and it says so before you agree: the restore points go with the environment, so there is no export afterwards, no restore, and no undo — not by us either. Nothing survives.
Both choices are yours, both are confirmed on the same screen, and neither waits for anyone here.
Locks stop a deletion
An environment can be locked, and so can a single application inside it. While a lock is armed, deletion does not proceed, and the review tells you exactly which lock is holding it and where to clear it. That is deliberate: a lock is what stops a deletion nobody wants, instead of a queue of people approving it.
Once it is gone
After the data is gone it cannot be recovered by extending retention, and it cannot be undone by support. That is why the confirmation is explicit and why it names what it destroys.
Getting your data out
What you can take yourself depends on whether the environment is still serving:
- A database copy, while the environment serves. A Company Admin can request a copy of one database from Hosting → Databases → Create backup now and download it before it expires. It is a new copy of that database at the moment you ask, not a historical restore point. See Backups and restores.
- One application's reset or restore, while the environment serves. That is a reviewed operation on a single application, and it leaves the application's siblings alone. See Backups and restores and Restores.
- Your files, and anything after service has ended, are arranged by email. There is no self-service download of the file archive, and no self-service restore of a whole environment. A copy of your files, or any restore once service has stopped, goes through support on the same path as any other restore — see Backups and restores. Ask before the recoverable-until date, because after it the restore points no longer exist.
Next
- Get a copy of your data: Backups and restores.
- Drive it from a terminal: Command line.
- Ask us something first: Support.