Cloudflare D1 Backup

This script creates an backup/export of a Cloudflare D1 SQLite database. It uses the D1 HTTP API to query for table definitions and data, then outputs SQL commands to recreate the database as-is.

This script has only been tested on small databases (~700KB). Please report any bugs using GitHub Issues.

Based on nora-soderlund/cloudflare-d1-backups, which requires to be ran inside a Worker. This repository uses the D1 HTTP API.

Usage

To create a backup, you must obtain:

• Your Cloudflare account ID. This can be found as the ID in the URL on the dashboard after dash.cloudflare.com/, or in the sidebar of a zone.
• Your Cloudflare D1 database ID. This can be found on the D1 page.
• Your Cloudflare API key. This can be created under the user icon in the top-right under "My Profile", then "API Tokens" in the sidebar. Make sure to have D1 write access (the script does not write to your database).

CLI

This will create the backup at backup.sql.

CLOUDFLARE_D1_ACCOUNT_ID=... CLOUDFLARE_D1_DATABASE_ID=... CLOUDFLARE_D1_API_KEY=... \
npx @cretezy/cloudflare-d1-backup backup.sql

The CLI also supports reading from .env.

You may also pass the --limit to add a LIMIT clause for each SELECT query. Default is 1000. You may need to lower if D1 crashes due to Isolate Has exceeded Memory Size. You can increase to speed up exports.

Library

npm i @cretezy/cloudflare-d1-backup

import { createBackup } from "@cretezy/cloudflare-d1-backup";

const backup = await createBackup({
  accountId: "...",
  databaseId: "...",
  apiKey: "...",
  // Optional, see note above on --limit
  limit: 1000,
});

backup will be the string of the backup commands.

Restoring a backup

npx wrangler d1 execute <database> --file=<backup.sql>

<database> must be the ID or name of the D1 database.