Retention Policies & Automated Storage Cleanup

RUSEON Core features an automated background retention scheduler (internal/recorder/) that continuously monitors disk utilization, enforces data retention policies, and prevents storage exhaustion.

---

Retention Strategies

RUSEON Core supports two complementary retention triggers configured in config.yaml:

1. Time-Based Retention (retention_days)

• Enforces maximum storage duration per camera or globally.
• Recordings older than retention_days (calculated from directory date and file epoch) are deleted on the next retention cycle.
• Example: Setting retention_days: 30 guarantees recordings beyond 30 days are automatically purged.

2. Capacity-Based Quota (min_disk_free_gb / max_disk_usage_percent)

• Prevents storage volumes from reaching 100% capacity, which would freeze database writes or crash the OS.
• If free disk space drops below min_disk_free_gb (e.g. 20 GB) or usage exceeds max_disk_usage_percent (e.g. 90%), the retention engine immediately triggers emergency FIFO pruning, deleting the oldest hourly recording directories until target free space is restored.

---

The Pruning Algorithm

1. Chronological Traversal (FIFO): The cleaner navigates the Storage File Layout starting from the oldest {YYYY}/{MM}/{DD}/{HH} directory.
2. Atomic Deletion: Deletes both the .mp4 video container and the accompanying .idx sidecar file.
3. Empty Folder Pruning: Automatically removes emptied hour, day, and month directories to keep the filesystem clean.
4. Database Sync: Removes deleted file records from the BadgerDB index so the REST API and UI timeline update immediately.
5. Prometheus Metrics: Emits ruseon_retention_deleted_bytes_total and ruseon_retention_deleted_files_total.

---

Configuration Reference

recording:
  enabled: true
  storage_path: "/data/recordings"
  
  # Time retention limit in days (0 = unlimited, bounded only by disk space)
  retention_days: 14

  # Emergency disk space safety thresholds
  min_disk_free_gb: 25
  max_disk_usage_percent: 90

  # Frequency of the cleanup cycle
  cleanup_interval: "10m"

Related Documentation

• Archive Overview
• Storage File Layout
• Crash Recovery
• Prometheus Metrics Reference