📦 appinsights-pino-logger

A lightweight, production-ready TypeScript logger built on Pino, designed for modern distributed applications. It provides:

⚡ Fast, pretty, colorized logging
🧵 Automatic correlationId tracking via AsyncLocalStorage
☁️ Optional Azure Application Insights integration
🔗 Inline single-line log formatting (no JSON metadata blocks)
🔧 Runtime configuration through logger.init()

Ideal for microservices, Kafka consumers, API gateways, and distributed systems needing reliable request tracing.

🚀 Features

| Feature | Description | | ---------------------------------- | ---------------------------------------- | | ⚡ Fast Pino logging | Pretty output, timestamps, colorized | | 🧵 AsyncLocalStorage correlationId | Automatic context propagation | | ☁️ Optional Azure AI | Only used if installed + enabled | | 🧩 Multiple log arguments | logger.info("a", 1, { b: 2 }) | | 🔧 Runtime configuration | serviceName, version, log level, AI keys | | 🧼 Clean inline logs | No multi-line metadata in console | | 🔄 Severity mapping | Translates Pino → Azure AI severity |

📥 Installation

npm install appinsights-pino-logger

Optional (only if you want Azure Telemetry):

npm install applicationinsights

🔧 Initialization

import { logger } from "appinsights-pino-logger";
// const { logger } = require("appinsights-pino-logger"); // CommonJS

logger.init({
  serviceName: "billing-service",
  version: "2.1.0",
  level: "debug",
  connectionString: process.env.APPINSIGHTS_CONNECTION_STRING,
  enableAI: true
});

Supported init() Options

| Option | Type | Description | | -------------------- | -------- | --------------------------- | | serviceName | string | Name of your service | | version | string | Version tag for logs | | level | string | Log level (default: info) | | timestamp | function | Custom timestamp formatter | | connectionString | string | Azure AI connection string | | instrumentationKey | string | Legacy Azure AI key | | enableAI | boolean | Toggle AI logging |

⚙️ Environment Variables (Optional)

SERVICE_NAME=service-api
SERVICE_VERSION=1.0.0
LOG_LEVEL=debug

APPINSIGHTS_CONNECTION_STRING=InstrumentationKey=xxxx...
ENABLE_APPINSIGHTS=true

🧱 Basic Usage

import { logger } from "appinsights-pino-logger";

logger.info("Service started");
logger.debug("Debug details");
logger.error("Something went wrong");

console.log = logger.log; // optional override console.log

🧵 Correlation ID

Automatic (Async Safe)

import { withContext, logger } from "appinsights-pino-logger";

withContext({ correlationId: "order-999" }, async () => {
  logger.info("Processing order");
  await new Promise(r => setTimeout(r, 200));
  logger.info("Order completed");
});

🔍 Log Output Format

Terminal output:

[2025-12-09 03:55:27] INFO: [correlationId: "test-corr-1232"][version: "1.2.0"] Payment created {"amount":100} USD {"userId":50}

✨ Why this log format?

🔹 One clean line per log entry
🔹 Optimized for local development readability
🔹 No bulky multi-line metadata blocks
🔹 Complete metadata still captured by Azure AI

☁️ Viewing Logs in Azure Application Insights

Basic:

traces
| order by timestamp desc

Filter by correlationId:

traces
| where customDimensions.correlationId == "order-999"

🧪 Logging Multiple Arguments

logger.info(
  "Payment created",
  { amount: 100 },
  "USD",
  { userId: 50 }
);

🧰 Express Example

import express from "express";
import { withContext, logger } from "appinsights-pino-logger";
import { v4 as uuid } from "uuid";

logger.init({ serviceName: "express-api", version: "1.0.0" });

const app = express();

app.use((req, res, next) => {
  const correlationId = req.headers["x-correlation-id"] || uuid();
  withContext({ correlationId }, next);
});

app.get("/hello", (_, res) => {
  logger.info("Request received");
  res.send("Hello");
});

app.listen(3000, () => logger.info("Server running"));

🧬 Kafka Example

consumer.on("message", msg => {
  withContext({ correlationId: msg.headers.correlationId }, () => {
    logger.info("Message received", msg.value);
  });
});

🧹 Graceful Shutdown

process.on("beforeExit", () => {
  logger.aiClient?.flush();
});

📝 License

MIT