Retry site backups

View as MarkdownOpen in Claude
Re-queues failed backups. The server already retries a failed copy on its own up to four times during its first day, so retry by hand only when the user asks. Body parameters: - `backups` (integer array, required): exact backup ids, 1 to 50, from the listing with `filter[phase][]=failed`. Answers **202** with the re-queued `site-backups` documents and `meta.skipped`, each row with `site_id`, `site_name`, `backup_id`, `reason_code` and `message`: `UNSUPPORTED` (the copy is not in phase `failed`), `SITE_UNREACHABLE` (website not connected), `CONFLICT` (another backup or a restoration running on that website), `PERMISSION_DENIED` (id outside the token's reach). Follow the result with `GET /site-backups?filter[id][]=...` and read `phase`. - **409**: the storage quota is exhausted and no overage add-on is active - nothing is retried. - **422**: `backups` missing, empty, over 50 ids, or containing an id that resolves to no row. - **403**: a read-only token. The MCP twin is the `site-backups-retry` tool.

Authentication

AuthorizationBearer

Personal access token created in the Modular DS dashboard; read-only tokens can only call GET endpoints.

Request

This endpoint expects an object.
backupslist of doublesOptional

Response

202 Accepted - one copy re-queued, one skipped

datalist of objectsOptional
jsonapiobjectOptional
metaobjectOptional