Week of Mar 28, 2025

Improvements

  • Prevent deployments of sites not ready to be deployed
  • Slack authorization handling

Fixes

  • Circle invitation validation
  • Aikido integration correctly matches repositories
  • Log files can now be cleared again
  • Daemons can now be created with startsecs=0
Week of Mar 17, 2025

Improvements

  • Updated font on invoice

Fixes

  • No notification when server failed to delete at provider
  • Supplying empty username when creating a site via the API
  • Aikido setup when using custom GitLab
  • Required email field when editing backup configuration
  • Server disappearing from dashboard when deleting a load balanced site
  • pool.d file missing when changing PHP version on an isolated site
  • stopwaitsecs missing in deamons API (was part of docs)
  • Removing Laravel specific daemons sometimes removed wrong deamon
  • Servers not marked as connecting while refreshing status
Week of Mar 3, 2025

Fixes

  • Creating an isolated site via the API with invalid usernames
Week of Feb 17, 2025

Improvements

  • Role ARNs are required when editing AWS credentials

Fixes

  • Reverb and Octane daemons are created with custom ports
  • “Run Job Now” correctly reports the status of the job
Week of Feb 10, 2025

Improvements

  • Improved performance of server and site name searching
  • Improved server password generation

Fixes

  • Notification channels are now correctly removed when transferring servers
  • Reverb site name is now updated when a site is renamed
  • Blackfire is now correctly removed for PHP 8.4 when the integration is disabled
  • Clean up failed Reverb installations
Week of Feb 3, 2025

Improvements

  • Clarify Deployment Script’s reloading of PHP FPM
  • Increased security requirements for new passwords

Fixes

  • Correctly handle transferring of AWS servers when using Role ARN based credentials
  • Fixed deleting of Circles with pending invitations
  • Fixed error when transferring servers that are part of a circle
Week of Jan 27, 2025

Improvements

  • SSH banners are now muted when Forge connects to a server
  • Newly provisioned servers now write a .hushlogin file to prevent errors caused by MOTD
  • Improved handling of disconnected servers when running scripts
  • Enabled X-Frame-Options=SAMEORIGIN header for improved security

Fixes

  • When deleting AWS servers, the EBS block volumes are also deleted
  • Reload PHP-FPM when restarting Nginx
  • Fixed create server modal not hiding in some cases
  • Improved authorization retry logic when connected to Bitbucket
Week of Jan 20, 2025

Improvements

  • Add support for Swoole with PHP 8.4
  • Add Thailand (ap-southeast-1) to AWS regions list
  • Add Mexico (mx-central-1) to AWS regions list
  • New AWS credentials will use short-lived tokens via AWS Roles

Fixes

  • Sites can now be safely renamed back to default
  • Add -s flag when syncing MySQL databases to prevent errors
  • Fixed recipe reports
Week of Jan 13, 2025

Improvements

  • Server names now include a wider variety of names
  • SSH keys generated by Forge will now be use ed25519
  • Node 22.x is now installed by default (previously 18.x)
  • Forge now detects the process status of all daemons
  • Server Ubuntu versions are now shown in the information bar

Fixes

  • Disabled Microsoft Teams notifications
  • Firewall IPs may be comma separated including spaces
  • Worker information will now show all known information
  • Daemon status syncing is now more resilient
  • Pinned awscli to v2.22.35 to resolve issues when using DigitalOcean and Vultr S3 providers with database backups
Week of Jan 6, 2025

Improvements

  • Add validation message when archiving balancers from API
  • Switched field order when installing existing certificates
  • Report rate limiting errors to the frontend

Fixes

  • Allow installing repositories with special characters in branch names
  • AWS VPC is no longer always required when creating new AWS servers
  • “Skip this step” button now redirects correctly when onboarding
  • Disabled “Blogo” starter kit within Statamic integration
  • Let’s Encrypt certificates should timeout less frequently
  • Forge now reloads PHP-FPM when reloading Nginx configurations (previously calling restart could cause moments of downtime)

Was this page helpful?