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

# Postgres

> Configure a Postgres destination for the GoldRush Pipeline API to batch-insert structured blockchain data.

The Postgres destination writes structured pipeline output to a PostgreSQL database using batch inserts. It follows the same batching model as the ClickHouse destination with defaults tuned for PostgreSQL workloads.

## Configuration

```yaml theme={null}
destination:
  type: "postgres"
  url: "postgresql://host:5432/db"
  user: "${PG_USER}"
  password: "${PG_PASSWORD}"
  batch_size: 1000
  flush_interval_ms: 5000
```

### Fields

| Field               | Type   | Required | Default | Description                   |
| ------------------- | ------ | -------- | ------- | ----------------------------- |
| `url`               | string | yes      | --      | JDBC connection URL           |
| `user`              | string | yes      | --      | Database user                 |
| `password`          | string | yes      | --      | Database password             |
| `batch_size`        | int    | no       | 1,000   | Rows per batch insert         |
| `flush_interval_ms` | long   | no       | 5,000   | Max time (ms) between flushes |

## How It Works

1. Rows are buffered in memory as they arrive from the pipeline.
2. When the buffer reaches `batch_size` **or** `flush_interval_ms` elapses since the last flush, a batch `INSERT` executes against the target table.
3. The `INSERT` SQL is dynamically generated from the field names present in each row.
4. All column names are automatically quoted in the generated SQL to handle PostgreSQL reserved words such as `user`, `from`, and `order`.