Logging System

DBackup uses three distinct logging systems:

1. System Logger - For application-wide logging (errors, debug info, operations)
2. Execution Logs - Structured logs for backup/restore job tracking (UI display)
3. Notification Logs - Records every notification sent for audit trail and preview rendering

System Logger

Added in v0.9.4-beta

The centralized logging system was introduced to replace scattered console.log calls throughout the codebase.

Overview

The system logger provides consistent, level-based logging across the entire application.

Location: src/lib/logger.ts

Basic Usage

import { logger } from "@/lib/logger";

// Simple logging
logger.info("Backup started");
logger.debug("Processing file", { filename: "backup.sql" });
logger.warn("Connection slow", { latency: 500 });
logger.error("Operation failed", { operation: "upload" }, error);

Child Loggers

For component-specific logging, create a child logger with context:

import { logger } from "@/lib/logger";

const log = logger.child({ service: "BackupService" });

// All logs will include { service: "BackupService" }
log.info("Starting backup job", { jobId: "abc123" });
// Output: { level: "info", service: "BackupService", jobId: "abc123", message: "Starting backup job" }

Log Levels

Level	Usage	When to Use
debug	Detailed debugging info	Development, troubleshooting
info	Normal operations	Important state changes
warn	Non-critical issues	Degraded functionality
error	Failures	Exceptions, failed operations

Environment Configuration

Control log output via the LOG_LEVEL environment variable:

# .env
LOG_LEVEL=debug   # Show all logs (development)
LOG_LEVEL=info    # Default (production)
LOG_LEVEL=warn    # Only warnings and errors
LOG_LEVEL=error   # Only errors

Output Formats

Development (colored, human-readable):

[2026-02-05T10:30:00.000Z] INFO  [BackupService] Starting backup job { jobId: "abc123" }

Production (JSON, machine-parseable):

{"timestamp":"2026-02-05T10:30:00.000Z","level":"info","service":"BackupService","message":"Starting backup job","jobId":"abc123"}

Error Handling Integration

The logger integrates with the custom error system:

import { logger } from "@/lib/logger";
import { wrapError, AdapterError } from "@/lib/errors";

const log = logger.child({ adapter: "mysql" });

try {
  await connectToDatabase();
} catch (error) {
  // wrapError() converts unknown errors to DBackupError
  log.error("Connection failed", { host: config.host }, wrapError(error));
  throw new AdapterError("mysql", "Failed to connect to database");
}

Best Practices

Do

• Use child loggers with context (service, adapter, step)
• Include relevant metadata as the second parameter
• Use appropriate log levels
• Use wrapError() for error logging

Don't

• Don't use console.log, console.error, etc. directly
• Don't log sensitive data (passwords, keys, tokens)
• Don't log inside hot loops (performance impact)

---

Middleware Logging

The Next.js middleware automatically logs API requests and security events.

Location: src/middleware.ts

API Request Logging

All API requests (except silent paths) are logged with timing information:

INFO  API request {"module":"Middleware","method":"GET","path":"/api/jobs","duration":"12ms","ip":"127.0.0.1"}
INFO  API request {"module":"Middleware","method":"POST","path":"/api/backup/run","duration":"45ms","ip":"192.168.x.x"}

Silent Paths (not logged to reduce noise):

• /api/health - Health check endpoint (frequent polling)
• /api/auth/get-session - Session validation (every request)

Rate Limit Event Logging

When a client exceeds the rate limit, a warning is logged:

WARN  Rate limit exceeded {"module":"Middleware","ip":"192.168.x.x","path":"/api/jobs","method":"POST","limiter":"mutation"}

Limiter Types:

Limiter	Applies To	Limit
auth	/api/auth/sign-in	Strict (prevent brute force)
api	GET/HEAD requests	Standard
mutation	POST/PUT/DELETE/PATCH	Stricter

IP Anonymization

For privacy compliance (GDPR), IP addresses are anonymized in logs:

Original IP	Logged As
192.168.1.100	192.168.x.x
10.0.0.50	10.0.x.x
127.0.0.1	127.0.0.1 (localhost unchanged)
::1	::1 (IPv6 localhost unchanged)
2001:db8::1	2001:db8:x (IPv6 prefix only)

---

Custom Error Classes

Location: src/lib/errors.ts

DBackup provides a hierarchy of custom error classes for consistent error handling:

Error Hierarchy

DBackupError (base)
├── AdapterError       - Database/storage adapter failures
├── ConnectionError    - Network/connectivity issues
├── ConfigurationError - Invalid config or settings
├── ServiceError       - Business logic failures
├── NotFoundError      - Resource not found
├── ValidationError    - Input validation failures
├── PermissionError    - RBAC authorization failures
├── AuthenticationError - Login/session failures
├── BackupError        - Backup operation failures
├── RestoreError       - Restore operation failures
├── EncryptionError    - Encryption/decryption failures
└── QueueError         - Job queue failures

Creating Custom Errors

import { AdapterError, BackupError, wrapError } from "@/lib/errors";

// Adapter-specific error
throw new AdapterError("mysql", "Connection timeout after 30s");

// Backup operation error
throw new BackupError("Dump failed: insufficient permissions");

// Wrapping unknown errors
try {
  await riskyOperation();
} catch (e) {
  throw wrapError(e); // Converts to DBackupError
}

Utility Functions

import {
  isDBackupError,
  getErrorMessage,
  getErrorCode,
  withContext
} from "@/lib/errors";

// Type guard
if (isDBackupError(error)) {
  console.log(error.code); // e.g., "ADAPTER_ERROR"
}

// Safe message extraction
const message = getErrorMessage(unknownError);

// Add context to errors
throw withContext(error, { jobId: "123", attempt: 2 });

---

Execution Logs (Job Tracking)

For backup and restore operations, DBackup uses structured execution logs that are displayed in the UI.

The LogEntry Structure

Location: src/lib/core/logs.ts

export interface LogEntry {
  timestamp: string;      // ISO 8601 format
  level: LogLevel;        // 'info' | 'success' | 'warning' | 'error'
  type: LogType;          // 'general' | 'command'
  message: string;        // Short, human-readable message
  stage?: string;         // Current execution stage
  details?: string;       // Long output (stdout, stack traces)
  context?: Record<string, any>; // Additional metadata
}

export type LogLevel = 'info' | 'success' | 'warning' | 'error';
export type LogType = 'general' | 'command';

Log Levels

Level	Usage	UI Color
info	Normal progress messages	Blue
success	Completed steps	Green
warning	Non-fatal issues	Orange
error	Failures	Red

Log Types

Type	Usage	UI Display
general	Status messages	Normal text
command	Shell commands, SQL	Monospace, collapsible

Usage in Services

The runner pipeline uses execution logs for job tracking:

Log Buffer Pattern

class BackupRunner {
  private logs: LogEntry[] = [];
  private currentStage: string = 'Initialization';

  private log(
    message: string,
    level: LogLevel = 'info',
    type: LogType = 'general',
    details?: string
  ) {
    const entry: LogEntry = {
      timestamp: new Date().toISOString(),
      message,
      level,
      type,
      stage: this.currentStage,
      details,
    };
    this.logs.push(entry);
    this.flushLogs(); // Periodic DB update
  }

  private setStage(stage: string) {
    this.currentStage = stage;
    this.log(`Starting ${stage}`, 'info');
  }
}

Example Usage

// Simple info message
this.log('Download started');

// Success with stage
this.setStage('Upload');
this.log('File uploaded successfully', 'success');

// Command with output
this.log(
  'Executing mysqldump',
  'info',
  'command',
  `mysqldump --host=db.example.com --user=backup mydb`
);

// Error with details
this.log(
  'Connection failed',
  'error',
  'general',
  error.stack
);

// Warning
this.log(
  'Slow connection detected',
  'warning',
  'general',
  `Latency: ${latencyMs}ms`
);

Execution Stages

Standard stages used throughout the pipeline:

Stage	Description
Initialization	Loading configuration, resolving adapters
Dump	Creating database dump
Compression	Applying GZIP/Brotli
Encryption	Encrypting with vault key
Upload	Transferring to storage
Retention	Cleaning up old backups
Completion	Final cleanup, notifications

For restore operations:

Stage	Description
Initialization	Loading configuration
Download	Fetching backup file
Decryption	Decrypting if encrypted
Decompression	Extracting if compressed
Restore	Applying to database
Verification	Optional integrity check
Completion	Cleanup

Log Persistence

Flushing Strategy

Logs are buffered in memory and flushed to the database periodically:

private async flushLogs() {
  // Debounced flush every 500ms
  await db.execution.update({
    where: { id: this.executionId },
    data: { logs: JSON.stringify(this.logs) },
  });
}

Database Storage

model Execution {
  id        String   @id
  logs      String   // JSON string of LogEntry[]
  // ...
}

Retrieving Logs

const execution = await db.execution.findUnique({
  where: { id: executionId },
});

const logs: LogEntry[] = JSON.parse(execution.logs || '[]');

Frontend Rendering

Stage Grouping

The UI groups logs by stage for better readability:

function ExecutionLogs({ logs }: { logs: LogEntry[] }) {
  const grouped = groupBy(logs, 'stage');

  return (
    <div>
      {Object.entries(grouped).map(([stage, entries]) => (
        <StageSection key={stage} name={stage}>
          {entries.map((log) => (
            <LogLine key={log.timestamp} entry={log} />
          ))}
        </StageSection>
      ))}
    </div>
  );
}

Command Collapsing

Command-type logs show the command itself, with expandable details:

function LogLine({ entry }: { entry: LogEntry }) {
  if (entry.type === 'command') {
    return (
      <Collapsible>
        <CollapsibleTrigger>
          <code>{entry.message}</code>
        </CollapsibleTrigger>
        <CollapsibleContent>
          <pre>{entry.details}</pre>
        </CollapsibleContent>
      </Collapsible>
    );
  }

  return <p className={levelStyles[entry.level]}>{entry.message}</p>;
}

Live Progress

For real-time updates during execution:

// Server: Update execution with progress
await db.execution.update({
  where: { id },
  data: {
    logs: JSON.stringify(logs),
    progress: {
      stage: currentStage,
      percent: calculatePercent(),
      message: lastLog.message,
    },
  },
});

// Client: Poll for updates
const { data } = useSWR(
  `/api/executions/${id}`,
  fetcher,
  { refreshInterval: 1000 }
);

Notification Logs

Added in v0.9.9-beta

Notification logging was introduced for audit trail and adapter-specific preview rendering on the History page.

Overview

Every notification sent through the system (per-job and system-wide) is logged to the NotificationLog table. This provides a full audit trail and enables adapter-specific preview rendering on the History page.

Service: src/services/notification-log-service.ts

Recording Notifications

Logging happens transparently in both dispatch points:

import { recordNotificationLog } from "@/services/notification-log-service";

// After sending a notification
await recordNotificationLog({
  eventType: "BACKUP_SUCCESS",
  channelId: channel.id,
  channelName: channel.name,
  adapterId: "discord",
  status: "success",           // or "error"
  title: payload.title,
  message: payload.message,
  fields: JSON.stringify(payload.fields),
  color: payload.color,
  renderedPayload: JSON.stringify(discordEmbed),
  executionId: execution.id,   // optional, for per-job notifications
  error: null,                 // error message if send failed
});

Key design: recordNotificationLog() is fire-and-forget — it catches and swallows all errors to never block notification delivery.

Dispatch Points

Location	Context
src/lib/runner/steps/04-completion.ts	Per-job backup notifications
src/services/system-notification-service.ts	System-wide events (login, restore, config backup, storage alerts)

Rendered Payloads

Each log entry stores adapter-specific rendered content for History page preview:

Adapter	renderedPayload Content	renderedHtml
Discord	Embed object (title, description, fields, color)	—
Slack	Block Kit blocks array	—
Teams	Adaptive Card body	—
Telegram	Parsed HTML message	—
Email	—	Full rendered React email HTML
Others	—	—

Data Retention

Notification logs are automatically cleaned by the "Clean Old Data" system task:

• SystemSetting key: notification.logRetentionDays
• Default: 90 days
• Configurable: 7 days to 5 years (Settings → General → Data Retention)