sabaoth backup create

Requests an on-demand backup of a service. Asks for confirmation unless you pass --yes, because running it twice queues two backups.

sabaoth backup create <service-id> [--name <name>] [--yes] [--org <org-id>] [--json]

Arguments

string

required

The ID of the service to back up, for example clu_abc.

Flags

string

A name for the backup, such as pre-upgrade. Use lowercase letters, numbers and hyphens only. If left out, the server picks a name.

boolean

default:"false"

Skip the confirmation prompt. Required when there’s no terminal.

string

Organization the service belongs to, instead of the selected one.

boolean

default:"false"

Print the result as JSON on stdout.

Examples

sabaoth backup create clu_abc
sabaoth backup create clu_abc --name pre-upgrade
sabaoth backup create clu_abc --name pre-migration --yes --json

Confirmation

Create a backup in Acme (org_abc):
  Cluster:  clu_abc
  Name:     pre-upgrade

Backup creation is queued, not immediate, and repeating it queues duplicate work.
Continue? [y/N]

Only y continues. Anything else cancels with DECLINED (exit 2). If there’s no terminal and you didn’t pass --yes, the command fails with CONFIRMATION_REQUIRED (exit 2) and sends no request.

Output

The command returns as soon as the backup is queued, not when it finishes:

{
  "id": "bak_abc",
  "service_id": "clu_abc",
  "name": "pre-upgrade",
  "status": "scheduled",
  "backup_type": "on-demand",
  "created_at": "2026-01-01T00:00:00Z",
  "expires_at": "2026-02-01T00:00:00Z"
}

Check on progress with sabaoth backup list.

Backup statuses

Status	Meaning
`scheduled`	Queued and waiting to start.
`running`	In progress.
`completed`	Finished and ready to restore from.
`failed`	The backup didn’t succeed.
`expired`	Past its retention period.

Behavior

The organization check happens on the server. Unlike postgres delete, this command doesn’t look up the service first, so it sends a single request. A service in another organization is reported as CLUSTER_NOT_FOUND.
Restoring a backup is done in the dashboard. See Backups.

If the result is unknown

If the request was sent but no response came back, the backup may or may not have started. The CLI never retries, to avoid starting a second one. It fails with UNKNOWN_MUTATION_OUTCOME. Run sabaoth backup list <service-id> to check before trying again.

Errors

Code	Exit	Meaning
`DECLINED`	2	You didn’t confirm. No backup was requested.
`CONFIRMATION_REQUIRED`	2	No terminal to ask for confirmation. Re-run with `--yes`.
`CLUSTER_NOT_FOUND`	1	No service with this ID in the effective organization.
`BACKUP_CONFLICT`	1	A backup may already be running, or the service can’t take one yet. Check with `sabaoth backup list`.
`BACKUP_NOT_PERMITTED`	1	The server refused. Common causes, in order: your email isn’t verified, your role can’t start backups, the organization has no active subscription, or backups aren’t allowed for this service.
`BACKUP_REQUEST_REJECTED`	1	The API rejected the request, for example because the name isn’t valid.
`INVALID_CLUSTER_ID`	1	The API rejected the value as a service ID.
`UNKNOWN_MUTATION_OUTCOME`	1	No response. See above.

Get Started

Account

Organizations

PostgreSQL

Backups

Reference

Arguments

Flags

Examples

Confirmation

Output

Backup statuses

Behavior

If the result is unknown

Errors

Arguments

Flags

Examples

Confirmation

Output

Backup statuses

Behavior

If the result is unknown

Errors

Related