VectorStoreMigration
API-facing view of a project's vector store migration.
API-facing view of a project's vector store migration.
| Property | Type | Required | Description |
|---|---|---|---|
project_id | string | Yes | A 24-character hexadecimal MongoDB ObjectId string. |
status | MigrationState | No | Default: "idle". |
target | MigrationTarget | null | No | |
attempt_id | string | null | No | Identifies the current start-to-cutover/cancel attempt, freshly generated by start_migration and carried by every dispatched re-embedding message. retry reuses it - it resumes the same attempt rather than starting a new one. Lets a message from an attempt that was cancelled (and possibly replaced by a new one with the same target_generation) be told apart from a message belonging to the live attempt, even though both would otherwise read as status == MIGRATING. |
vector_store_ids | string[] | No | The vector stores this migration's cutover/cancel applies to, fixed at start time. A vector store created after start (writes aren't blocked for creation) never got a target-generation collection or re-embedded files, so it must not be touched by cutover or cleanup. Default: []. |
target_generation | integer | No | The collection_generation every vector store in this project will have after cutover.Since all of a project's vector stores migrate together, this is uniform across the project: current_generation + 1. Default: 0. |
error | string | null | No | Set when the migration as a whole failed for a reason that isn't any single file's fault - e.g. an unexpected exception (DB unreachable, timeout) partway through cutover or cancel's per-store cleanup. Distinct from a VectorStoreMigrationFileInDb's own error. Cleared on the next start or successful retry. |
created_at | integer | No | |
updated_at | integer | No | |
finished_at | integer | null | No | |
id | string | Yes | The migration identifier, equal to the project identifier. |
We use cookies on our website. We use them to ensure proper functioning of the site and, if you agree, for purposes such as analytics, marketing, and targeting ads.