Automate database and file backups for October CMS to off-site storage — encrypted and restorable.

Demo URL: https://october-demo.renatio.com/backend/backend/auth/signin
Login: backup
Password: backup

The backup is a zip file with the directories you specify and a dump of your database, stored on any of the filesystems configured in October. The plugin cleans up old backups, monitors their health, notifies you when something goes wrong and restores the database from any backup it created.

• A backup database and application files with mouse click
• Restore the database from any backup with a console command
• S3, FTP, SFTP, Dropbox storage support
• MySQL, PostgreSQL, SQLite and Mongo databases support
• A configurable scheduler for automatic backups
• Backups triggered over HTTP, for hosting without a system cron
• Queued backups that survive execution time limits
• A history of every backup run, with filters, bulk delete and automatic retention
• Per-action permissions
• Automatic cleanup of old backups
• Extensive settings options
• Encryption and password protection
• Verification of every created archive
• Monitoring the health of all backups
• Mail, Discord and webhook notifications

This plugin requires PHP 8.3, with the ZIP module, October CMS 4.0 and Laravel 12.40 or higher. It's not compatible with Windows servers.

If you are using an older version of Laravel and October, take a look at one of the previous versions of this plugin.

The package needs free disk space where it can create backups. Ensure that you have at least as much free space as the total size of the files you want to backup.

mysqldump, pg_dump or mongodump must be installed for the database engine you back up.

Backup runs are serialized under a cache lock, which needs a cache store shared across processes: file, database or redis. With the array store backups still run, but may overlap.

Something that is free has little or no perceived value. Users do not commit to free products and only use them until something else looks nice and is free comes along. When I invest my time in the development of a new plugin I commit to supporting and maintaining it. I ask my customers to do the same. I do not make money from this plugin by advertisements, upgrades or additional services like hosting or setup.

Did you know that 30% of your purchase or donation goes to help fund the October Project?

My plugins take many hours to develop (40-120+) and even more hours to document and maintain. My paid plugins have to pay for both this time, and the time I am spending on free plugins and less successful paid plugins. This means that it will take even a successful plugin years to become profitable. Please consider buying an extended license if you want me to continue to maintain these plugins for the very small fee I ask in return or hire me for adding functionality that you feel is missing but valuable.

If you like this plugin, give this plugin a Like or Make donation with PayPal.

Please check my other plugins.

Please use GitHub Issues Page to report any issues with plugin.

Reviews should not be used for getting support or reporting bugs, if you need support please use the Plugin support link.

Icon made by Darius Dan from www.flaticon.com.

The plugin registers a Backups menu in the backend.

• Application backup creates a backup of all project files and the database. By default the project base path is backed up without vendor and node_modules; change that in the plugin Settings.
• Database backup creates a database-only backup.
• Clean old backups removes old backups according to the Cleanup settings.

Access is granted per action under Settings → Administrators, on the Backups tab. Super users bypass all of them.

Copy restore command only hides the button. Restoring runs from the console, so anybody with shell access can restore regardless of this permission.

Restoring replaces your database with the one stored inside a backup archive. It is available from the console only, because it cannot be undone. The terminal button in the backups list copies the matching command to your clipboard.

php artisan backup:restore
php artisan backup:restore --path="October CMS/2026-01-15-09-20-58.zip" --disk=local

Only the database is restored. Project files inside the archive are never touched.

You will be logged out. The dump contains backend_users and sessions, so accounts return to the state they were in when the backup was taken.

Restoring is not transactional. An import that fails halfway leaves the database inconsistent, which is why a safety backup is taken first - keep the path it prints.

A few things to know:

• Anything passed as --password shows up in ps and the shell history; on a shared server use BACKUP_RESTORE_PASSWORD or let the command ask.
• The safety backup is encrypted only with --password, never with the password from the settings, because that one lives in the database being replaced. Without --password it is stored in plain form - delete it once you have verified the restore.
• The restore waits for no other backup: if one is running, it stops before touching the database.
• Tables not present in the dump are left untouched.
• When the archive holds several dumps, the one matching --connection is used, by database name first and connection name second. If none matches, the restore stops and lists what the archive contains.
• PostgreSQL and SQLite dumps contain no drop statements, so they can only be restored into an empty database.
• The native client - mysql, psql or sqlite3 - must be in PATH, and proc_open enabled.

A safety backup taken without --password can be imported without this application:

unzip -p "storage/app/October CMS/safety-2026-01-15-09-20-58.zip" "db-dumps/*.sql" > /tmp/recover.sql
mysql -u USER -p DATABASE < /tmp/recover.sql

A Maximum execution time of ... seconds exceeded error means the backup is too large to finish in a single web request. You can:

1. Raise max_execution_time in your PHP configuration.
2. Use the Scheduler (recommended).
3. Use Console commands.
4. Enable Queued backups.

Enabling Queue backups on the Destination tab runs backend backups through the queue, so the request returns immediately and the backup continues in the background. This requires a running queue worker:

php artisan queue:work

Only a flash message is shown after clicking; open the History page to see the output once the job completed.

Leave this setting off if you do not run a queue worker, otherwise backups started from the backend will never run.

The History page records every backup and cleanup run: when it happened, how long it took, how large the archive was and whether it succeeded. Clicking an entry shows the output of that run.

The history is a journal of what the plugin did, not a list of the archives that exist - the backups list always reads the disk. Removing an entry, selected entries or the whole history never touches the archives. Entries expire according to Keep history entries for days in the settings.

Go to Settings and open Backup Manager under the Backup section.

Configure how often the plugin runs the database backup, application backup, cleanup and health monitor.

Important note: For scheduled tasks to operate correctly you must set up the scheduler: https://docs.octobercms.com/4.x/setup/scheduler.html

For hosting without a system cron, Allow external trigger publishes three addresses that run the tasks over HTTP:

https://example.com/backupmanager/trigger/<token>/db
https://example.com/backupmanager/trigger/<token>/app
https://example.com/backupmanager/trigger/<token>/clean

Point an external cron service (cron-job.org, EasyCron, an uptime monitor) at the one you need. A 200 means the task ran, a 500 means it failed.

The address is the password. Anybody who knows it can run the task, so keep it out of tickets, screenshots and anything else that gets shared. If it leaked, Generate new addresses replaces the token and every old address stops working at once. Switching the trigger off disables the addresses immediately.

Use HTTPS only - the token travels in the URL and ends up in server, proxy and cron service logs. When the calling service can send headers, call https://example.com/backupmanager/trigger/db (or app, clean) with an X-Backup-Token header instead.

The endpoint answers 404 to a wrong token, an unknown task and a disabled trigger alike, and limits requests to five per minute per IP address and ten per minute across all addresses combined. With Queue backups enabled it returns as soon as the task is queued, and only one job per task waits in the queue at a time.

Password-protects the archive. Use a long string and keep it safe - without it, you will never be able to open your backup. Leave it blank for an unprotected archive.

With the default settings the file backup contains the whole project, including .env with APP_KEY, database credentials and storage keys. Without an archive password anyone who can read the backup can read the secrets.

Either set an archive password, or add .env to the Exclude list if your hosting recreates it. The backups page warns as long as .env is backed up without a password.

Test decrypting an archive once, so you know the restore path works before you need it.

Warning: the default macOS archive utility cannot open encrypted ZIP files. Use The Unarchiver or BetterZip.

A backup is considered unhealthy when the latest backup is too old or the backups use more storage than allowed.

If mysqldump / pg_dump is not in a default location, or the dump exceeds the default 60 second timeout, configure the connection in config/database.php:

'connections' => [
	'mysql' => [
		'driver' => 'mysql'
		...,
		'dump' => [
		   'dump_binary_path' => '/path/to/the/binary', // only the path, so without `mysqldump` or `pg_dump`
		   'use_single_transaction',
		   'timeout' => 60 * 5, // 5 minute timeout
		   'exclude_tables' => ['table1', 'table2'],
		   'add_extra_option' => '--optionname=optionvalue',
		]
	],

By default backups are saved to storage/app/October CMS/. Create a dedicated disk (for example backups) in config/filesystems.php and select it in the plugin Settings.

Supported storage drivers: local, S3, FTP, SFTP and Dropbox. Configure the disk in config/filesystems.php, then select it on the Destination tab of the plugin Settings.

composer require league/flysystem-aws-s3-v3 "^3.0"

A sample s3 disk is included in config/filesystems.php.

composer require league/flysystem-ftp "^3.0"

'ftp' => [
    'driver' => 'ftp',
    'host' => env('FTP_HOST'),
    'username' => env('FTP_USERNAME'),
    'password' => env('FTP_PASSWORD'),

    // Optional FTP Settings...
    // 'port' => env('FTP_PORT', 21),
    // 'root' => env('FTP_ROOT'),
    // 'passive' => true,
    // 'ssl' => true,
    // 'timeout' => 30,
],

composer require league/flysystem-sftp-v3 "^3.0"

'sftp' => [
    'driver' => 'sftp',
    'host' => env('SFTP_HOST'),

    // Settings for basic authentication...
    'username' => env('SFTP_USERNAME'),
    'password' => env('SFTP_PASSWORD'),

    // Settings for SSH key based authentication with encryption password...
    'privateKey' => env('SFTP_PRIVATE_KEY'),
    'password' => env('SFTP_PASSWORD'),

    // Optional SFTP Settings...
    // 'hostFingerprint' => env('SFTP_HOST_FINGERPRINT'),
    // 'maxTries' => 4,
    // 'passphrase' => env('SFTP_PASSPHRASE'),
    // 'port' => env('SFTP_PORT', 22),
    // 'root' => env('SFTP_ROOT', ''),
    // 'timeout' => 30,
    // 'useAgent' => true,
],

Install the Dropbox Adapter plugin and configure the disk as described in its documentation.

• backup:run - run a backup; --only-db backs up only the database
• backup:clean - clean old backups
• backup:list - display the status of all monitored destination filesystems
• backup:monitor - check the health of all monitored destination filesystems
• backup:restore - restore the database from a backup archive, see Restoring a backup

Notifications go out by e-mail, to Discord and to any service accepting a webhook. A channel is used once it has an address; leave a field empty to disable it. Configure them on the Notifications tab of the plugin Settings.

A notification is sent when a backup or cleanup succeeds or fails, and when the monitor finds a healthy or unhealthy backup.