For the complete documentation index, see llms.txt. This page is also available as Markdown.

Step 3. Generate

Use this step to review the job summary, tune performance settings, and start generating.

1

Review the job summary

Review the included and excluded tables.

Double-check privacy coverage before you start based using the PII icon.

Job summary showing included/excluded tables and privacy status
2

Tune generation settings (optional)

Generation settings mainly affect speed, memory usage, and database load.

If you read from a production-like source, prefer smaller batches and fewer connections.

3

Start generating and monitor the job

Select Generate.

If you need logs or performance info, use:

Job summary and generation settings

The job summary is your final check.

It confirms:

  • Which tables are included or excluded.

  • Which tables contain sensitive data.

  • Whether sensitive columns are protected by generators.

Advanced generation settings

Batch sizes are defined in datapoints, not rows.

Datapoints are calculated as:

rows × columns

Example:

10,000 rows × 100 columns = 1,000,000 datapoints (1M)

You can tune these settings:

  • Batch size The maximum datapoints Syntho reads, generates and writes per batch. Default: 1M. Value: integer. Increase it to speed up reads and writes. Memory usage goes up. Decrease it if you see memory pressure or timeouts.

  • Maximum number of connections The maximum parallel connections used for writing. More connections can speed up large jobs. It can also overload the database. When reading from production-like sources, keep this conservative.

  • Simultaneous read/write When enabled, the next batch is generated while the previous batch is still being written. Disabling this will make generation slower but uses less memory.

Troubleshooting

SQL parameter limit exceeded

Some database drivers cap the number of parameters per SQL statement.

Large batch size can exceed that cap.

Example error:

Fix

Reduce batch size.

Start by reducing 1M → 0.5M.

If it persists, reduce again.

The exact limit depends on your database and driver.

Last updated

Was this helpful?