Backups & restore

Infrastructure → Backups (/gravity/backups) manages automated backup policies for your cloud instances - what gets backed up, on what schedule, and for how long. A separate area, Backup Validations, records restore-drill results - read the limits in Backup validation before treating it as recovery evidence.

Backup policies

A backup policy has a name, a provider scope (AWS, Azure, GCP, or all providers), a schedule written as a cron expression, a retention period in days, and an active/inactive toggle. The policy list shows each one’s schedule, retention, how many instances it covers, its status, and when it was last applied.

Create Policy asks for a name, provider, schedule (for example 0 2 * * * for daily at 02:00 UTC), and retention in days. You can edit any of those fields, plus the active toggle, from the policy’s detail page afterward.

Status	Meaning
Active	Applied and current
Pending	Apply is in progress
Error	The last apply attempt failed
Inactive	Turned off - not currently applying

A policy’s detail page also shows Attached Instances - the instances the policy currently covers.

Caution

There’s currently no way to add or remove instances from a policy’s coverage from this screen. If you need instances attached to or detached from a policy, contact support.

Applying a policy

Apply Now is the action that actually configures automated snapshots at your cloud provider - AWS DLM, Azure Recovery Services, or GCP resource policies, depending on the policy’s provider. Creating or editing a policy in Sencai only stores the policy; nothing is configured at your provider until you apply it. The policy’s status reflects the result.

Applying requires the policy to already cover at least one instance, so the platform knows which cloud account to configure - see the note above if a policy has no attached instances yet.

Applying a policy requires the Admin or Owner role. A Member, Auditor, or Viewer sees the Apply Now button but gets a permission error when they use it.

Caution

Applying a policy to your provider isn’t switched on in every environment yet. If Apply Now returns a permission error and you hold Admin or Owner, the dispatch path hasn’t been enabled for your account - ask your account team. Until then a policy is a stored record only; configure the snapshot schedule directly in your provider’s console.

Deleting a policy

Deleting removes the policy record in Sencai only. It does not remove any snapshot schedule already configured at your provider - if you delete a policy you had previously applied, also remove the corresponding backup schedule directly through your provider’s console if you don’t want it running any more.

Restoring

There’s no in-app “restore this instance” action today. Restoring an instance from one of its snapshots is done directly through your cloud provider’s own console or tooling. An instance’s detail page shows how many restore points your provider currently reports for it, but that’s a count, not a picker - use it to confirm snapshots exist, not to restore from here.

Backup validation

Operations → Backup Validations (/gravity/operations/backup-validations) lists restore-drill records for your organization - what was validated, when, whether the restore test passed, and how long it took.

Caution

Treat this screen as a log of drills that have been run, not as proof your backups are restorable. No recurring drill schedule is running today. Restore drills are currently implemented for AWS RDS instances only - any other provider records an explicit “not implemented for this provider” result, never a pass. The Trigger Drill button does not currently work from the browser. Talk to your account team before relying on this screen as recovery evidence.

See Operations for the rest of what that section covers.

What’s next

• Operations - backup validation and the rest of the operational picture
• Instances - the resources backup policies protect
• Connect a cloud provider
• Audit log - records of policy changes and applies