Skip to content

offline-mode.md

Latest commit

 

History

History
150 lines (114 loc) · 3.59 KB

offline-mode.md

File metadata and controls

150 lines (114 loc) · 3.59 KB

Offline Mode

WhiteBoxAI SDK supports offline mode for robust operation when network connectivity is unreliable.

Overview

Offline mode provides:

  • Persistent Queue: SQLite-based storage that survives application restarts
  • Auto-Sync: Background synchronization every 60 seconds (configurable)
  • Priority-Based: Operations are synced based on priority (CRITICAL > HIGH > NORMAL > LOW)
  • Retry Logic: Automatic retry with exponential backoff (max 3 attempts)
  • Thread-Safe: Supports concurrent operations safely

Quick Start

from whiteboxai import WhiteBoxAI

# Enable offline mode with auto-sync
client = WhiteBoxAI(
    api_key="your-api-key",
    enable_offline=True,
    offline_dir="./whiteboxai_offline",
    offline_auto_sync=True,
    offline_sync_interval=60
)

Configuration

Basic Configuration

client = WhiteBoxAI(
    api_key="your-api-key",
    enable_offline=True,              # Enable offline mode
    offline_dir="./offline_queue",    # Storage directory
    offline_max_queue_size=10000,     # Max queued operations (0 = unlimited)
)

Auto-Sync Configuration

client = WhiteBoxAI(
    api_key="your-api-key",
    enable_offline=True,
    offline_auto_sync=True,           # Enable background sync
    offline_sync_interval=60,         # Sync every 60 seconds
)

Managing the Queue

Check Queue Status

status = client.get_offline_status()
print(f"Queue size: {status['queue_size']}")
print(f"Last sync: {status['last_sync_time']}")
print(f"Failed operations: {status['failed_count']}")

Manual Sync

result = client.sync_offline_queue()
print(f"Synced: {result['synced']}")
print(f"Failed: {result['failed']}")

Cleanup Old Operations

# Remove operations older than 7 days
client.cleanup_offline_queue(older_than_days=7)

Priority Levels

Set priority for operations:

# HIGH priority prediction
monitor.log_prediction(
    inputs={"feature1": 1.0},
    output={"prediction": 0.85},
    priority="HIGH"
)

# CRITICAL priority for important events
monitor.log_prediction(
    inputs={"feature1": 1.0},
    output={"prediction": 0.95},
    priority="CRITICAL"
)

Priority levels: CRITICAL, HIGH, NORMAL (default), LOW

Best Practices

  1. Storage Location: Choose a persistent location for offline_dir
  2. Queue Size: Set appropriate offline_max_queue_size based on your storage
  3. Sync Interval: Balance between freshness and API load
  4. Cleanup: Regularly cleanup old operations to prevent queue bloat
  5. Monitoring: Monitor queue size and failed operations

Error Handling

Offline mode handles errors gracefully:

try:
    monitor.log_prediction(inputs=data, output=prediction)
except Exception as e:
    # Operation is automatically queued if offline mode is enabled
    print(f"Logged to offline queue: {e}")

Production Deployment

For production deployments:

client = WhiteBoxAI(
    api_key=os.getenv("WHITEBOXAI_API_KEY"),
    enable_offline=True,
    offline_dir="/var/lib/whiteboxai/queue",
    offline_max_queue_size=50000,
    offline_auto_sync=True,
    offline_sync_interval=30,  # More frequent sync in production
)

# Monitor queue health
import schedule

def check_queue_health():
    status = client.get_offline_status()
    if status['queue_size'] > 10000:
        alert_ops_team(f"Queue size: {status['queue_size']}")

schedule.every(5).minutes.do(check_queue_health)

See Also