RabbitMQ message profiling

This tutorial shows how to profile RabbitMQ messages consumed with @golevelup/nestjs-rabbitmq. Each consumed message produces a profile — with a Message panel plus any HTTP, cache, or database activity the handler triggered — that appears in a dedicated RabbitMQ table at /_profiler.

Prerequisites

• @eleven-labs/nest-profiler installed and configured
• @golevelup/nestjs-rabbitmq consuming messages via @RabbitSubscribe

Step 1 — Install the package

pnpm add @eleven-labs/nest-profiler-rabbitmq

@golevelup/nestjs-rabbitmq and amqplib are optional peer dependencies — the ones you already use to consume messages.

Step 2 — Register the module

Add RabbitMqCollectorModule to the application that consumes your messages (the same process that hosts the profiler):

import { Module } from '@nestjs/common';
import { ProfilerModule } from '@eleven-labs/nest-profiler';
import { RabbitMqCollectorModule } from '@eleven-labs/nest-profiler-rabbitmq';
import { RabbitMQModule } from './rabbitmq.module';

@Module({
  imports: [
    ProfilerModule.forRoot({ isGlobal: true }),
    RabbitMqCollectorModule.forRoot(),
    RabbitMQModule,
  ],
})
export class AppModule {}

Step 3 — Keep your consumer unchanged

You do not change anything for profiling — write an ordinary @RabbitSubscribe handler:

import { Injectable } from '@nestjs/common';
import { RabbitSubscribe } from '@golevelup/nestjs-rabbitmq';
import type { ConsumeMessage } from 'amqplib';

@Injectable()
export class NarrationService {
  @RabbitSubscribe({
    exchange: 'articles.events',
    routingKey: 'published.*',
    queue: 'tts.narration',
  })
  async createGeneration(message: ArticleEvent, raw: ConsumeMessage): Promise<void> {
    // … your business logic …
  }
}

Step 4 — Inspect the profile

Publish a message (or trigger your normal flow) and open /_profiler. Messages appear in a dedicated RabbitMQ table — separate from the HTTP/GraphQL and Commands tables — and you can isolate them with the Type filter (choose RabbitMQ). Open one to find:

• Message — exchange, routing key, handler, redelivered flag, consumer/delivery tags, headers and payload (the HTTP request/response tabs are hidden)
• HTTP Client, Database, … — whatever the handler triggered

A handler that throws is captured too: its profile is marked failed (status 500) and the thrown error appears in the Exceptions tab.

Try it in the example app

The example API wires this end to end behind a feature flag (off by default). Start a broker and run it with the flag on:

docker compose up -d rabbitmq
FEATURE_RABBITMQ=true pnpm example:dev

Then publish a message and open /_profiler to see it profiled:

curl -X POST http://localhost:3000/notifications \
  -H 'content-type: application/json' \
  -d '{ "subject": "Deploy done", "body": "v1.2.3 shipped" }'

Configuration

RabbitMqCollectorModule.forRoot({
  enabled: true, // set false to disable per environment
  captureHeaders: true, // AMQP headers (sensitive ones masked)
  captureBody: true, // deserialized payload — disable if payloads are large
  maskHeaders: ['x-tenant-secret'], // merged with the built-in mask list
});

How it works

A consumed message has no HTTP request/response, so the module registers an IContextAdapter for the rmq execution context that creates a fresh profile per message with a rabbitmq entrypoint (entrypoint.type = 'rabbitmq', the message details on entrypoint.data). The core ProfilerInterceptor opens a CLS context, runs the handler — so profile-scoped collectors keep capturing — then runs all collectors and persists the profile. The module also registers the rabbitmq entrypoint type, contributing the RabbitMQ list table, the Message detail tab and the rabbitmq option on the Type filter. @golevelup/nestjs-rabbitmq is an optional peer dependency: when no consumer runs, the module never produces a profile.