> ## Documentation Index
> Fetch the complete documentation index at: https://docs.e6data.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Database

> database

ReferenceApplies to: Unified engine deployments

`[database]`

Unified uses SlateDB at `<storage_root>/metadata/<service>/`. The backend and base prefix are computed. See [storage](https://docs.e6data.com/transformation/operations/configuration/storage.md).

## Settings

### `flush_interval`

Sets how often SlateDB groups pending record writes before making them durable.

* **Default:** Unset
* **Type:** duration; optional

> **Example**
>
> `{ secs = 0, nanos = 100000000 }` chooses a 100 ms interval. Only set this field with SlateDB; it is invalid for Memory or SQLite metadata.

### `migration_batch_size`

Limits how many saved records one migration step updates together.

* **Default:** `256`
* **Type:** integer; positive

> **Example**
>
> `256` migrates at most 256 records in one atomic batch before saving progress for the next batch.

### `payload_version`

Selects the application record format for an explicit metadata migration.

* **Default:** Unset
* **Type:** integer; optional

> **Example**
>
> Version `1` refers to an engine metadata format, not an Iceberg or SQL table version. Change it only as part of a planned metadata migration.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.