Back up tenants

EIDGuard backs up every onboarded External ID tenant on a schedule, and you can start a backup at any time. Each backup of a tenant produces a recovery point: a read-only copy of the tenant's configuration that you can browse, compare and restore from.

How backups run#

• Backups run as Azure Automation jobs in your subscription. They sign in to each tenant with its read-only backup app, so a backup cannot change the tenant.
• One schedule serves all tenants. Each tenant has a recovery point objective (RPO): how old its newest recovery point may become. The deployment has a default RPO, and each tenant can override it. Each scheduled run backs up the tenants that are due.
• Recovery points are stored in the storage account in your deployment's resource group and protected from deletion for at least one day. See Protect recovery points from deletion.

Configure the backup schedule#

You need the Admin role. Other roles see the schedule as read-only.

1. Open Configuration, then the Protection tab.
2. Under Backup schedule, set:
   • Default RPO: Every hour, Every 2 hours, Every 4 hours, Every 6 hours, Every 8 hours, Every 12 hours or Daily (the default).
   • Time zone: Not set (UTC), or a time zone. See Time zone.
   • Run hour: the hour the daily backup runs. With an RPO shorter than a day this field is called Anchor hour, and backups run at that hour and every RPO interval after it.
3. Select Save schedule. A line under the settings describes when backups will run.

Tenants use the default RPO unless one is set for them on the Tenants page.

Time zone#

The time zone is one deployment setting with two effects:

• Schedules run in it. Backups, the daily verification and, with private networking, the worker VM's start and stop window follow the selected zone. The run hour is a local hour and stays at that local time through daylight saving changes.
• The dashboard shows times in it, including next runs, job times, recovery points and drift reports.

With Not set (UTC), schedules run in UTC and each viewer sees times in their browser's time zone.

Changing the time zone keeps the hour number you chose (2 stays 2, now 2:00 AM local time) and moves the backup and verification schedules to that zone. Backups are never left unscheduled during the change. If part of the change does not complete, the page says so and selecting Save schedule again finishes it.

On the two nights a year the clocks change, one interval between backups is an hour longer and one is an hour shorter. A local time that does not exist that night runs at the equivalent moment just after the change; a local time that happens twice backs each tenant up once. The Overview's 30-day backup calendar always counts UTC days.

Use IANA time zone names, for example America/New_York or Asia/Kolkata (not the older alias Asia/Calcutta).

Run a backup now#

You need the Operator role.

• On Recovery points, select Run backup, or
• on Operations, select Start backup in the Backup row.

Either action backs up every onboarded tenant immediately, whether or not it is due, and does not change the schedule. Take a backup before a planned configuration change so you have a recovery point from just before it.

Browse recovery points#

Recovery points lists every recovery point, newest first, with Created, Tenant, Origin, Age and Objects. Filter by tenant or origin, or search.

Open a recovery point to see:

• Captured, Objects, Retained until and Deletion protection.
• Last differential: the most recent comparison of this recovery point, with View report.
• Captured resources: each resource type and its object count. Select Inspect to list the objects, and open an object to see its captured configuration as read-only JSON.
• Reports: comparison and restore reports for this recovery point.

From a recovery point you can Compare it with the live tenant (Operator), Restore from it (Operator), or Export its manifest as a JSON file.

What a recovery point contains#

A recovery point is a folder named for the tenant and the capture time, holding one JSON file per backed-up resource type and a manifest that lists them. Comparison and restore reports are added to it later. For the resource types included, see Supported resource types.

Retention#

Retention decides how long recovery points are kept. You need the Admin role to change it.

1. Open Configuration > Protection.
2. Under Retention, enter the number of days and select Apply.

• The default is 90 days. Enter 0 to keep recovery points forever. The maximum is 3650 days.
• The minimum is shown next to the field. It is at least 30 days and never shorter than the deletion protection window.
• A tenant can have its own retention; see Per-tenant protection.
• Retention applies to existing recovery points too: those older than the new value are deleted once their deletion protection window has passed.

Recovery points of offboarded tenants#

When you offboard a tenant, its recovery points are kept. They stay listed with offboarded · points retained until retention removes them, and you can browse them. Comparing or restoring needs the tenant onboarded again, because offboarding deleted the apps EIDGuard uses to sign in to the tenant. For how long they are kept, see Recovery points of an offboarded tenant.

Verify that backups are running#

• Overview > Protected tenants shows whether each tenant has a recovery point inside its RPO.
• The backup reliability card shows the last 30 days of backup jobs. Select a day to open its job.
• Jobs lists every backup job with its status and log.
• The bell icon (Needs attention) lists overdue tenants and failed jobs.

If a backup job fails, the failed-jobs alert emails the alert address. See Monitoring and alerts.

Related pages#

• Manage protected tenants
• Protect recovery points from deletion
• Restore from a recovery point
• Verify backups and detect drift