@kb-labs/adapters-pino-http

Part of KB Labs ecosystem. Works exclusively within KB Labs platform.

Pino HTTP transport for streaming logs to KB Labs REST API.

Features

• ✅ Batching - Configurable batch size and flush interval
• ✅ Retry Logic - Exponential backoff on HTTP failures
• ✅ Graceful Shutdown - Flushes pending logs before exit
• ✅ Error Handling - Logs to stderr if HTTP fails
• ✅ Lightweight - Minimal dependencies, uses native fetch

Installation

pnpm add @kb-labs/adapters-pino-http

Usage

Basic Setup

import pino from 'pino';

const logger = pino({
  transport: {
    target: '@kb-labs/adapters-pino-http',
    options: {
      url: 'http://localhost:5050/api/v1/logs/ingest',
      batchSize: 50,
      flushIntervalMs: 3000,
    },
  },
});

logger.info('Hello from Pino HTTP Transport!');

With Platform DI (Recommended)

Configure in .kb/kb.config.json:

{
  "platform": {
    "adapters": {
      "logger": "@kb-labs/adapters-pino"
    },
    "adapterOptions": {
      "logger": {
        "level": "info",
        "options": {
          "transport": {
            "target": "@kb-labs/adapters-pino-http",
            "options": {
              "url": "http://localhost:5050/api/v1/logs/ingest",
              "batchSize": 50,
              "flushIntervalMs": 3000
            }
          }
        }
      }
    }
  }
}

Then use in code:

import { initPlatform } from '@kb-labs/core-platform';

const platform = await initPlatform();

platform.logger.info('This log will be sent to REST API!', {
  plugin: 'my-plugin',
  executionId: 'exec-123',
});

Configuration

| Option | Type | Default | Description | |--------|------|---------|-------------| | url | string | http://localhost:5050/api/v1/logs/ingest | REST API endpoint for log ingestion | | batchSize | number | 50 | Number of logs to batch before sending | | flushIntervalMs | number | 3000 | Max time in ms to wait before flushing batch | | retryAttempts | number | 3 | Number of retry attempts on HTTP failure | | retryDelayMs | number | 1000 | Initial retry delay in ms (exponential backoff) | | headers | Record<string, string> | {} | Custom HTTP headers (e.g., for authentication) | | debug | boolean | false | Enable debug logging to stderr |

How It Works

1. Pino logs are written to the transport stream
2. Batch accumulation - Logs are collected into a batch
3. Flush trigger - Batch is sent when:
   • Batch size reaches batchSize (immediate flush)
   • flushIntervalMs timer expires (delayed flush)
   • Process exits (graceful shutdown)
4. HTTP POST - Batch is sent to REST API with retry logic
5. Retry on failure - Exponential backoff (1s → 2s → 4s)
6. Error logging - Failures are logged to stderr

Performance

• Batching reduces HTTP overhead: 50 logs/request vs 50 requests
• Non-blocking: Uses async fetch, doesn't block Pino
• Memory-safe: Ring buffer in REST API prevents OOM

Debugging

Enable debug mode to see transport activity:

{
  "transport": {
    "target": "@kb-labs/adapters-pino-http",
    "options": {
      "url": "http://localhost:5050/api/v1/logs/ingest",
      "debug": true
    }
  }
}

Output:

[PinoHTTP] Flushing 50 logs to http://localhost:5050/api/v1/logs/ingest
[PinoHTTP] Flush successful
[PinoHTTP] Shutting down, flushing pending logs...
[PinoHTTP] Shutdown complete

License

KB Public License v1.1 - KB Labs Team