S3 Purge Local Backup Feature + Fallbacks when local files are purged

[ejscheepers]

This pull request introduces a comprehensive enhancement to the S3 backup workflow, delivering significant disk space savings, improved multi-connection organization, and seamless operational continuity when local backup files are purged.

---

Part 1: S3 Purge Local Backup After Upload

• New Configuration: Added s3_purge_local boolean setting, fully integrated across database, backend, and frontend.
• Behavior: When enabled, local .sql backup files are automatically deleted immediately after a successful upload to S3, freeing valuable server disk space.
• User Control: A new UI toggle has been added to the S3 Storage Settings page, clearly labeled "Purge Local Backup After Upload", giving administrators full control over storage behavior.

---

Part 2: Intelligent S3 Fallback for All Backup Operations

To ensure zero disruption when local files are purged, a robust fallback mechanism has been implemented across all backup operations:

• DownloadBackup(), RestoreBackup(), and CompareBackups() now:

  1. Check for the existence of the local backup file.
  2. If missing but an s3_object_key is stored:
     • Automatically download the file from S3 to a secure temporary location.
     • Proceed with the operation using the downloaded file.
     • Clean up the temporary file upon completion.
  3. If neither local file nor S3 key exists → return a clear, user-friendly error.

• New Helper: ensureBackupFileAvailable() in backup_service.go centralizes this logic for consistency and maintainability.

---

Part 3: Automatic S3 Folder Organization by Database Connection

Backups are now automatically organized into subfolders based on the sanitized connection name, transforming flat S3 structures into a clean, scalable hierarchy:

Before:

s3://my-bucket/backups/backup_2025-11-14.sql

After:

s3://my-bucket/backups/production-postgres/backup_2025-11-14.sql
s3://my-bucket/backups/staging-mysql/backup_2025-11-14.sql

• Backend Implementation:

  • New UploadFileWithPath() method in S3Storage accepts a subfolder.
  • New getObjectKeyWithPath() builds full S3 object keys using connection name.
  • uploadToS3IfEnabled() now passes sanitized connection name as subfolder.
  • All upload callers updated to include connection context.

• Key Benefits:

  • Easy identification of backups per database.
  • Enables per-connection S3 lifecycle policies (e.g., 30-day retention for prod, 7-day for dev).
  • Fully backward compatible — download/restore uses stored s3_object_key with full path.

---

Frontend Enhancements (TypeScript/React)

• New Field: s3_object_key?: string added to Backup interface in backup.ts.
• Visual Indicators:
  • Added blue cloud badge with S3 label and cloud icon.
  • Displayed in:
    • Backup History List (mobile & desktop)
    • Dashboard Recent Activity
  • Consistent styling across light/dark themes.

---

Key Features & Guarantees

Feature	Description
Space Saving	Enable purge to eliminate local storage bloat
Seamless Fallback	All operations work even with S3-only storage
Visual Clarity	Blue S3 badges instantly identify cloud backups
Scalability	Supports unlimited database connections
Backward Compatibility	No breaking changes; local workflows unchanged
Cleanup Safety	Temp files always removed post-operation
Error Resilience	Clear messages if backup is unavailable

---

Result for Users

Administrators can now:

1. Enable "Purge Local Backup After Upload" in S3 settings.
2. Upload backups to S3 → local file is automatically deleted.
3. Continue using Download, Restore, and Compare — system transparently pulls from S3.
4. See at a glance which backups are in S3 via blue cloud badges.
5. Save significant server disk space while maintaining full functionality.

[vercel]

@ejscheepers is attempting to deploy a commit to the Dendi Team on Vercel.

A member of the Team first needs to authorize it.

[ejscheepers]

Apologies for the AI PR Description, was just easier to summarise using it.

[dendianugerah]

Doesn’t matter man haha. I’ve already seen the code, looks really solid. I’ll test it out tomorrow

[ejscheepers]

Well done on the project by the way! I've been looking for a tool like this without bloat for a while now. Currently I use pg_back on the server with a cron job, but the lack of ui is frustrating.

[dendianugerah]

lgtm thanks

[ejscheepers]

Thanks for merging! Just checking — when is the next Docker image build/push scheduled, or does it happen automatically on merge to main?

[dendianugerah]

yeah it happens automatically when merged into main