Skip to content
plugboard.appRequest a demo

Let us host it

Support, restores and exits

Email [email protected]. That address is the front door for every school, self-hosted included, and it is the one to use for a restore request or anything urgent.

Include these four things and most issues are answerable on the first reply:

  1. Your version, from Settings, Licence & billing, under Deployment diagnostics.
  2. What you did, and what happened instead. Including the exact wording of any error.
  3. When it started, and what changed immediately before. This answers more cases than anything else on the list.
  4. Whether it affects everyone or one person. A single account behaving differently is usually a role or a module rather than a fault.

Included support runs 9 am to 5 pm Sydney time on weekdays; see support coverage. An earlier licence keeps the support its agreement names, along with any service credits.

Support cannot reach into your deployment until you let it. An administrator grants support access under Settings, Compliance for 24 hours, 3 days or 7 days, with a reason, and can end it at any time. Ask us what we need it for before you grant it; the reason you give is shown to support and kept in your audit log.

While access is on, support can:

  • read your deployment’s version and status, and your administrators’ names, addresses, last sign-in and two-factor state
  • list your backups
  • restore a backup, only when the grant includes restoring backups
  • email an administrator a link to set their password

Every request is recorded in your audit log before it reads anything, naming the support operator, and a refused request is recorded too.

Support never sees or passes on a setup link. When you ask for one, we request it while support access is on, and your deployment emails it to the address on the administrator’s account through your school’s own outgoing mail. It works once and stops working after four hours, and setting a password with it signs that account out everywhere else. A school with no outgoing mail connected cannot be sent one this way.

If nobody at the school can sign in to grant access, the principal or business manager sends us a written request from the school’s domain, and we confirm it by calling the school’s published number before acting.

Support access covers the product’s support channel only. We run the server your deployment is on, and hosting administration is separate from this switch.

Your deployment takes its own encrypted backups every night at 03:00 and keeps them for 90 days, from its first boot; see backups in the app. An administrator with backup.manage can restore one of your school’s backups from Settings, Backups without asking us.

To have us do it, email [email protected], grant support access with Include restoring backups ticked, and tell us:

  • Which deployment, if you have more than one.
  • The point in time you want to go back to.
  • Whether you want the whole school or a specific extraction. Restoring everything to yesterday discards today. Often the answer is one table, or one record, pulled out of a backup and put back.

We will confirm what will be lost before doing anything.

For self-hosted deployments, restores are yours. See backups and restore.

You do not need to ask. Settings, Backups has Export this school’s data, a portable copy of every record for your school, at any time.

Connector credentials in the export stay encrypted with your deployment’s SECRETS_MASTER_KEY. To use them elsewhere you need that key, which the school can ask us for as below; otherwise reconnect each connector with its own credentials after the import.

Both directions are supported and follow the same shape: a backup, a provision, a restore, a DNS change.

Managed to self-hosted:

  1. Ask in writing. The principal or business manager writes to [email protected] from the school’s domain. We call back on the school’s published number, which we look up ourselves, before anything moves. This is the same check as for a locked-out administrator, because what comes next is the key to your school’s data.
  2. We issue your self-hosted licence and setup bundle. You provision a server and install Plugboard. See install it yourself.
  3. We hand you your master key ring. Outside your managed server, its master key is held only in offline escrow, so we take it from there: every key your server has used, because a backup sealed before a rotation opens only with its own key. It comes by hand, in a file encrypted with a passphrase, and the passphrase is spoken on the confirmed call. It never goes through our panel or the support ticket, or in plain email. Before the first start, it replaces the generated key in .env, with the tenant ID we give you, so your server accepts your own backups.
  4. Your newest backup is restored on your server. Take a backup and verify it in Settings, Backups on the managed desk. A restore opens backups from its own server’s history or offsite storage, not a downloaded file, so we arrange the copy with you. Then sign in and check that connectors still test successfully. That last check is what proves the secrets decrypted.
  5. Rotate the master key on your server once the restore has worked, so the escrowed copy opens nothing your server writes from then on. Keep the old key in the ring until the backups it sealed age out. See rotating it.
  6. Add the new SSO redirect URI at your identity provider before cutover, keeping both registered during the change.
  7. DNS moves. We keep the old hostname resolving for a while so links already sent keep working.
  8. Once you confirm your desk is running, we close the managed deployment.
  9. Your licence continues, at the same price: self-hosting and managed hosting are priced alike. See plans.

Self-hosted to managed: the same in reverse, and easier, because we provision the target.

Between regions: the same again, within managed hosting. Tell us early. The DNS and identity provider changes have lead times that dwarf the data movement.

Remember the identity provider in every direction. Redirect URIs are absolute, and a cutover that changes the hostname without changing them locks everyone out of SSO.

Self-hosted deployments keep running. The licence check is local and offline, so there is no phone-home that can fail closed.

Managed customers can export at any time without our involvement.

Severity levels, what happens in the first fifteen minutes, and when customers and regulators are told are covered in incidents.

We tell you about anything affecting your deployment, with what we know at the time, and on more than one channel besides the status page.