@errpulse/node

Backend error monitoring SDK for Node.js, Express, and Next.js. Part of ErrPulse — the error monitoring tool that runs with one command.

Installation

npm install @errpulse/node

Quick Start

// Zero-config — auto-captures uncaught exceptions, rejections, console.error
import "@errpulse/node";

That's it. Errors are sent to http://localhost:3800 by default.

What Gets Caught

| Error Type | How | | ------------------------------ | -------------------------------------- | | Uncaught exceptions | process.on('uncaughtException') | | Unhandled promise rejections | process.on('unhandledRejection') | | Express route errors (4xx/5xx) | Error handler middleware | | Next.js API route errors | withErrPulse() wrapper | | console.error calls | Monkey-patch | | console.log/warn/info/debug | Monkey-patch (opt-in) | | Memory warnings | Periodic process.memoryUsage() check | | All HTTP requests | Request handler middleware |

Configuration

import { init } from "@errpulse/node";

init({
  serverUrl: "http://localhost:3800",
  projectId: "my-api",
  enabled: true,
  sampleRate: 1.0,
  captureConsoleErrors: true,
  captureConsoleLogs: false, // opt-in: capture console.log/warn/info/debug to Logs dashboard
  captureUncaughtExceptions: true,
  captureUnhandledRejections: true,
  monitorMemory: true,
  memoryThresholdMB: 512,
  memoryCheckIntervalMs: 30000,
  beforeSend: (event) => event, // Modify or drop events before sending
});

Config Reference

| Option | Type | Default | Description | | ---------------------------- | ---------- | ------------------------- | -------------------------------------------------- | | serverUrl | string | "http://localhost:3800" | ErrPulse server URL | | projectId | string | undefined | Project identifier for multi-project setups | | enabled | boolean | true | Enable or disable the SDK | | sampleRate | number | 1.0 | Sample rate from 0.0 to 1.0 (1.0 = capture all) | | captureConsoleErrors | boolean | true | Capture console.error calls | | captureConsoleLogs | boolean | false | Capture console.log/warn/info/debug to Logs view | | captureUncaughtExceptions | boolean | true | Capture uncaught exceptions | | captureUnhandledRejections | boolean | true | Capture unhandled promise rejections | | monitorMemory | boolean | true | Monitor memory usage and emit warnings | | memoryThresholdMB | number | 512 | Memory threshold in MB before warning | | memoryCheckIntervalMs | number | 30000 | How often to check memory (ms) | | beforeSend | function | undefined | Callback to modify or drop events before sending |

Express Integration

import express from "express";
import { init, expressRequestHandler, expressErrorHandler } from "@errpulse/node";

init({ serverUrl: "http://localhost:3800", projectId: "my-api" });

const app = express();
app.use(expressRequestHandler()); // Track all requests — must be first
// ... your routes ...
app.use(expressErrorHandler()); // Catch route errors — must be last

The request handler captures:

• HTTP method, URL, status code, and response duration
• Correlation ID (from X-ErrPulse-Correlation-ID header, or auto-generated)
• Request and response headers (sensitive headers redacted)
• Request and response bodies (capped at 16 KB)

Next.js Integration

import { withErrPulse } from "@errpulse/node";

export const GET = withErrPulse(async (req) => {
  const data = await db.query();
  return Response.json({ data });
});

Manual Capture

import { captureError, captureMessage } from "@errpulse/node";

captureError(new Error("Payment failed"), { userId: "123" });
captureMessage("Deployment started", "info", { version: "2.0" });

Graceful Shutdown

import { close } from "@errpulse/node";

process.on("SIGTERM", () => {
  close(); // Flushes buffered events and removes all listeners
  process.exit(0);
});

Error Correlation

When paired with @errpulse/react, the backend SDK reads the X-ErrPulse-Correlation-ID header injected by the frontend. The dashboard shows the full chain: user action -> frontend request -> backend error.

Documentation

• Full SDK Docs
• GitHub Repository

License

MIT