v4.0.0-beta.381

[andrasbacsai]

Changes

• feat/fix: Able to import full DB backups for PG/MySQL/MariaDB
• feat: Ability to set the initial root user (instance admin user) of Coolify in the .env file via environment variables (Read More):
  • ROOT_USERNAME=
  • ROOT_USER_EMAIL=
  • ROOT_USER_PASSWORD=
• feat: Registration is automatically disabled if you set up a root user via the .env file
• feat: Logic, error handling, and clean log output for each scenario of setting a root user in the .env file
• feat: Move Docker cleanup settings to their own tab
• feat: Added Docker cleanup executions UI with very detailed cleanup log and message. This will help with debugging storage problems.
• feat: Restore backup from server file
• feat: Add Infomaniak OAuth
• feat: Added a disk usage check frequency setting
• feat: Log Horizon worker name to app deployment jobs.
• feat: Add is_coolify_host to the server API responses.
• feat: Some new labels on all containers:
  • coolify.resourceName, coolify.serviceName, coolify.projectName, coolify.environmentName
• feat: New URL structure for better performance and fewer DB queries
• feat: Docker volume data can optionally be cloned alongside the resource (so it is a true full clone of the resource, everything is exactly the same (except the name ofc)) --> Disabled it for now, see reasons: fix/feat: Improves Cloning #4777 (comment).
• feat: Backup Retention Options (whichever limit is reached first will trigger the cleanup):
  • Number of backups to keep -> A number of backups to keep
  • Days to keep backups -> A number of days to keep backups
  • Maximum storage (GB) -> A number of GB or decimal of maximum allowed storage for this backup job
• feat: Configure commands for production
• feat: Configure modals for better error handling
• feat: Added Purify for better HTML sanitization
• feat: Add backup retention days, backup amount, and max allowed storage to S3
• feat: Add backup retention days and max allowed storage to locally stored backups
• feat: When deleting a backup schedule, you can now select to delete all backups from S3 as well
• feat: When deleting a single backup, you can now select to delete the backup from S3 as well
• feat: New backup retention UI.
• feat: Multi-delete on S3 and locally stored backups -> Now multiple backup files and backup executions are deleted together in parallel for improved performance and faster executions of deletion...
• Security fix: After initial installation, the Registration page is exposed with the registration form. This is usually not an issue as one installs and registers within seconds/minutes after the URL becomes publicly accessible. However, if you leave the registration page too long without registering a user, a malicious actor can potentially register and compromise your system.
• fix: Use finished_at instead of updated_at for the end time so that the time calculations are correct
• fix: Set finished_at as immutable so it cannot be changed later
• fix: Use wire:navigate whenever possible to make everything feel like a single-page application, and overall navigating Coolify now feels soooooo much smoother.
• fix: Only call removeOldBackups function in the DatabaseBackupJob.php if the backup is completed and successful; this caused problems before
• fix: Delete backup folder and parent folder if folders are empty when deleting local backups to clean up unused folders
• fix: Do not remove executions from DB until both S3 and local backups have been deleted and successfully processed; otherwise, backups will never be deleted from S3.
• fix: All routes now use /environment/{environment_uuid} instead of the environment name.

This helps because we can now use any character in environment names instead of being limited to URL-safe characters. Also, all other URLs are constructed with the uuid way, so we should do it consistently everywhere.

• fix: Nixpacks predefined variables are now used as build-args and could be overwritten from the defined env variables in Coolify.
• fix: Use local monaco-editor to remove reliance on Cloudflare CDN and improve performance (that's why we have +56K changed lines).
• fix: Use local dropzonejs
• fix: Always use Docker cleanup frequency no matter if threshold or force cleanup is enabled
• fix: Disable copy button if the context is not secure (http)
• fix: Cloning resource name + volume names
• fix: When having multiple Postgres init scripts, saving a change to one init script would delete all other init scripts.
• fix: Changing the file name of an existing init script would create a new init script instead of renaming the file.
• fix: Deleting an init script does not remove it from the server.
• fix: Remove custom-postgres.conf if input is empty or null as the file is not needed on the server anymore
• fix: Some UI fixes
• fix: API - Projects & Applications endpoints
• fix: Modal closing logic
• fix: Increase default PHP memory limit
• fix: Penpot health checks
• fix: NocoDB health check
• fix: Supabase template
• fix: Trigger with external DB
• fix: Make DB public for Supabase
• fix: CSRF bug for Label Studio
• fix: Better cloud jobs handling
• fix: Add lower timeout (ssh, docker) to jobs which are checking the servers, because it could cause problems if a high number of jobs are running.
• fix: Instance-wide GitHub apps are not available on other teams other than the source team
• fix: Undefined variable $fs_path error on file mounts for databases
• fix: SQLSTATE[42703]: Undefined column: 7 ERROR: column "additional_servers_count" of relation "applications" does not exist
• fix: Exclude computed count properties (additional_servers_count, additional_networks_count) loaded by global scope to prevent errors
• fix: Remove auto-generated properties (id, created_at, and updated_at) from replicate as that caused issues
• fix: Cloning a service to a separate server. To be able to clone a service to a new server, we need to set the server_id to the DB as well; otherwise, it will be cloned on the same server.
• fix: Volume naming after cloning
• fix: Project/environment-level cloning and also fixed resource-level cloning (same fixes described below for both)
• fix: Application cloning:
  • fix: Disable auto-generation of URLs and labels if read-only labels is false.
  • fix: Labels were not auto-generated for CloneMe (project/environment-level cloning) if read-only labels is true; now they are generated again
  • fix: Some settings get lost when cloning Applications because we need to clone the row in the application_settings table as well
  • fix: Clone file and directory mounts
  • fix: Clone tags
  • fix: Clone scheduled tasks
  • fix: Clone preview deployment settings
• fix: Databases cloning:
  • fix: Clone tags
  • fix: Clone volumes
  • fix: Clone file and directory mounts
  • fix: Clone backup schedules
• fix: Services cloning:
  • fix: Clone environment variables
  • fix: Clone tags
  • fix: Clone scheduled task
  • fix: Clone volumes
  • fix: Clone file and directory mounts
  • fix: Clone backup schedules for service databases
• fix: For read-only labels, a read-only Monaco editor is displayed
• fix: Render HTML on 500 error pages correctly
• fix: GitHub JWT token timing issue (more information about this issue here Expiration time' claim ('exp') is too far in the future probot/probot#1426).
  • issuedAt -1min -> this is to avoid token not yet valid issues
  • expiresAt +8min -> this is to make sure the token is not valid longer than 10 min (to avoid the issue described in the linked issue) as tokens have to have a lifespan of less than 10min, this should help with time differences of less than about 1min (most cases) - as the total valid time is 9min (with the -1min) -> for any case of over 50s time out of sync (System vs GitHub API time) it will throw an error.
  • fix: Expiration time' claim ('exp') must be a numeric value
• refactor: Dates and UI design of Backup executions, Scheduled Tasks executions, and Deployments executions tabs
  • redesigned the deployment executions tab
  • show full commit message with an expand button on the deployment executions tab
  • show only the first 7 characters of the commit hash, like on GitHub - makes the UI cleaner
  • move date transformation to a shared function
  • move duration calculation to a shared function
  • dates are now always displayed with their timezone next to them
  • remove dependency on dayjs (improves performance)
  • redesigned status badges with colors to make it easier to see the current status
  • redesigned execution log boxes so that they all have a similar look and feel and contain similar information
  • added start and end times for backups, scheduled tasks, deployments, and Docker cleanup executions
  • calculate the duration for backups, scheduled tasks, deployments, and Docker cleanup executions and added it to the UI
  • remove duplicate and unused code
• refactor: Rename functions to make code more clear
• refactor: Use camelCase consistently
• refactor: Consolidate code
• refactor: Better error handling in github.php and GithubAppPermissionJob.php
• refactor: Move things to separate functions
• refactor: Improve password security by not allowing compromised passwords
• refactor: Refactored the UI and DB to make it easier to use - 0 now means unlimited retention (so a value is now required for retention).
• refactor: Simplified some code
• refactor: Refactored some duplicated code
• refactor: Refactored code to be more centralized in one location
• chore: Remove unused authentik stuff
• chore: Refactored environment variable relationships
• chore: Refactored a lot of DB queries to improve performance
• chore: Upgraded all dependencies
• chore: Switched up read-only labels checkbox to make more sense -> Read-only labels are now the default and mean that they are read-only (not editable), and Coolify will autogenerate the labels. If you uncheck read-only, you can edit them, and Coolify will not autogenerate any labels.
• chore: Renamed number_of_backups_locally to database_backup_retention_amount_locally
• chore: Use native Livewire functions instead of scripts in some cases.

Issues

• fix [Bug]: Project environment path is not encoded when the name includes a forward slash #2461
• fix [Bug]: Saving changes to database initialization script deletes the other initialization scripts #3273
• fix [Bug]: Domain is removed everytime we call Update Application API #4420
• fix [Bug]: Can't deploy a Python Poetry project #4582
• fix [Bug]: [N8N] Timezone not being followed #4603
• fix [Bug]: LanceDB: Invalid ENV Settings - AnythingLLM (Includes fix) #4658
• fix [Bug]: Editing cronjobs returns 500 #3760
• fix [Bug]: Unable to restore with the file from Backup all databases in coolify #3919
• fix [Bug]: Supabase database expose #3297
• fix [Bug]: Supabase Public schema not publicly available  #2861
• fix [Bug]: Supabase: An error occurred while setting up the SSL connection #3050
• fix [Bug]: Backup page not working #4682
• fix [Bug]: Destination creation doesn't work #4416
• fix [Bug]: System Wide GitHub Apps does not work #2643
• fix [Bug]: Add file mount in storage result "Undefined variable $fs_path" #4668
• fix [Bug]: Clone supabase (created from templates) to another server #3571
• fix [Bug]: Cloning WordPress service doesn't work to other server, and doesn't clone some vars #4050
• fix [Bug]: Clone skips domains and proxy config #4073
• fix [Bug]: Clone Environment / Project causes SQLSTATE[42703] error #4592
• fix [Bug]: Problem while Cloning Environment into another Server #4572
• fix [Bug]: Cloned project: delete dialog asks old project name #4713
• fix [Bug]: Cloning applications into other servers not working #4725
• fix [Enhancement]: Delete old backups #4782
• fix [Bug]: UI refreshes webpage on navigation on Macos Firefox #4615
• fix [Bug]: claim ('exp') is too far in the future #4702
• fix [Bug]: Deployment finished a day ago in <time> is wrong #3313
• improves [Bug]: Disk Usage after a build on builder server stays high (>70%) #4775

---

New Navigation

new-navigation.mp4

Old Navigation

old-navigation.mp4

New Deployment Logs:

Old Deployment Logs:

Docker Cleanup Logs:

Backup Retention:

Root user environment variables do not pass validation rules

Root user created successfully

Root user already exists, skipping creation

[andrasbacsai]

For the NodeBB service, plugins installed via the admin dashboard are uninstalled when the service is restarted.

NodeBB/NodeBB#12958

Can you please open an issue for this? It is not so simple to fix.