Separate a placement group with a restart
POST /api/v2/placement-groups/{id}/actions/resolve
Fix a placement group that automatic separation could not fix. Hostup shuts down one VPS that shares hardware with another set, moves it to separate hardware and starts it again. That VPS is offline for a short time.
Call it only when the group has needsAttention: true and actions.canResolve.allowed: true in GET /api/v2/placement-groups/{id}. Send acceptRestart: true to confirm the restart. A contact (sub-account) needs access to the VPS that will restart.
Returns 202 with the group. The VPS that moves shows status: migrating. Poll GET /api/v2/placement-groups/{id}/status until health is ok.
Authentication: An API key with the scope write:vm. About keys and scopes
Request
Path parameters
idstringrequiredpg_…).Body
acceptRestartbooleanrequiredtrue. Confirms that one VPS restarts.Example
curl -X POST "https://cloud.hostup.se/api/v2/placement-groups/pg_01hxa3b4c5d6e7f8g9h0j1k2m3/actions/resolve" \ -H "Authorization: Bearer $HOSTUP_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "acceptRestart": true }'Responses
202 The move is queued. The group is reconciling while it runs.
{ "id": "pg_01hxa3b4c5d6e7f8g9h0j1k2m3", "name": "Production database", "policy": "spread", "health": "reconciling", "needsAttention": false, "compatibilityKey": null, "members": [ { "vpsId": "vps_01hxa3b4c5d6e7f8g9h0j1k2m3", "name": "db-primary", "instanceUrl": "/instance/vps_01hxa3b4c5d6e7f8g9h0j1k2m3", "placementSet": "a", "status": "placed", "addedAt": "2026-06-28T10:02:00.000Z" }, { "vpsId": "vps_01hxa3b4c5d6e7f8g9h0j1k2m4", "name": "db-replica", "instanceUrl": "/instance/vps_01hxa3b4c5d6e7f8g9h0j1k2m4", "placementSet": "b", "status": "migrating", "addedAt": "2026-06-28T10:03:00.000Z" } ], "actions": { "canRename": { "allowed": true, "reason": null }, "canDelete": { "allowed": true, "reason": null }, "canAddMember": { "allowed": true, "reason": null }, "canResolve": { "allowed": false, "reason": "not_needed" } }, "createdAt": "2026-06-28T10:00:00.000Z", "updatedAt": "2026-06-28T10:05:00.000Z"}Response fields
idstringpg_…).namestringpolicystringspread. One of spread.healthstringok: every set is apart. reconciling: a VPS is being moved. violation: VPSes from different sets share hardware. One of ok, reconciling, violation.needsAttentionbooleantrue when automatic separation could not finish. Check actions.canResolve.compatibilityKeystring or nulldeprecatednull.membersarray of objectvpsIdstringvps_…).namestringinstanceUrlstringplacementSetstringa, b, c.statusstringplaced: on hardware no other set uses. migrating: being moved. colocated: shares hardware with a VPS from another set. unknown: its location could not be read. One of placed, migrating, colocated, unknown.addedAtstringactionsobjectcanRenameobjectallowedbooleanreasonstring or nullnull when it is.canDeleteobjectallowedbooleanreasonstring or nullnull when it is.canAddMemberobjectallowedbooleanreasonstring or nullnull when it is.canResolveobjectPOST /api/v2/placement-groups/{id}/actions/resolve can separate the group now.allowedbooleanreasonstring or nullnot_needed: nothing to fix with a restart. no_restart_target: no separate hardware is free right now. null when allowed is true. One of not_needed, no_restart_target.createdAtstringupdatedAtstringErrors
Errors are Problem Details with a stable code. Branch on the code, not on the text.
| 400 | code is invalid_request. acceptRestart is missing or not true. An unknown field in the body also returns it. |
| 401 | Unauthorized. Authentication is required. |
| 403 | You are signed in as a contact without access to the VPS that would restart. code is contact_permission_required. |
| 404 | No placement group with this ID in your account. code is placement_group_not_found. |
| 409 | Nothing to resolve now. code is resolve_not_available. Either the group needs no fix, a move is already running or no separate hardware is free. |
| 429 | Rate limited. Retry after the limit resets. 429 responses include Retry-After seconds plus X-RateLimit-* headers. |
| 500 | Internal error. Retry later or contact support if the issue persists. |