TypeScript icon, indicating that this package has built-in type declarations

3.7.1 • Public • Published

NPM Version

MultiversX NestJS Microservice Monitoring Utilities

This package contains a set of utilities commonly used for monitoring purposes in the MultiversX Microservice ecosystem. The package relies on Prometheus to aggregate the metrics and it is using prom-client as a client for it.


sdk-nestjs-monitoring is delivered via npm and it can be installed as follows:

npm install @multiversx/sdk-nestjs-monitoring


The package exports performance profilers, interceptors and metrics.

Performance profiler

PerformanceProfiler is a class exported by the package that allows you to measure the execution time of your code.

import { PerformanceProfiler } from '@multiversx/sdk-nestjs-monitoring';

const profiler = new PerformanceProfiler();
await doSomething();
const profilerDurationInMs = profiler.stop();

console.log(`doSomething() method execution time lasted ${profilerDurationInMs} ms`);

The .stop() method can receive two optional parameters:

  • description - text used for default logging. Default: undefined
  • log - boolean to determine if log should be printed. If log is set to true, the logging class used to print will be Logger from "@nestjs/common". Default: false
import { PerformanceProfiler } from '@multiversx/sdk-nestjs-monitoring';

const profiler = new PerformanceProfiler();
await doSomething();
profiler.stop(`doSomething() execution time`, true);

The output of the code above will be "doSomething() execution time: 1.532ms"

Cpu Profiler

CpuProfiler is a class exported by the package that allows you to measure the CPU execution time of your code. Given that JavaScript is a single-threaded language, it's important to be mindful of the amount of CPU time allocated to certain operations, as excessive consumption can lead to slowdowns or even blockages in your process.

import { CpuProfiler } from '@multiversx/sdk-nestjs-monitoring';

const profiler = new CpuProfiler();
await doHttpRequest()
const profilerDurationInMs = profiler.stop();

console.log(`doHttpRequest() method execution time lasted ${profilerDurationInMs} ms`);

The .stop() method can receive two optional parameters:

  • description - text used for default logging. Setting the description automatically triggers the printing of the PerformanceProfiler value. Default: undefined
import { CpuProfiler } from '@multiversx/sdk-nestjs-monitoring';

const httpReqCpuProfiler = new CpuProfiler();
await doHttpRequest();
httpReqCpuProfiler.stop(`doHttpRequest() execution time`);

const cpuProfiler = new CpuProfiler();
await doSomethingCpuIntensive();
cpuProfiler.stop(`doSomethingCpuIntensive() execution time`);

The output of the code above will be

doHttpRequest() execution time: 100ms, CPU time: 1ms doSomethingCpuIntensive() execution time: 20ms, CPU time 18ms

Note that a big execution time does not necessarily have an impact on the CPU load of the application. That means that, for example, while waiting for an HTTP request, the JavaScript thread can process other things. That is not the case for CPU time. When a method consumes a lot of CPU time, Javascript will not be able to process other tasks, potentially causing a freeze until the CPU-intensive task is complete.


The package provides a series of Nestjs Interceptors which will automatically log and set the CPU and overall duration for each request in a prom histogram ready to be scrapped by Prometheus.

LoggingInterceptor interceptor will set the execution time of each request in a Prometheus histogram using performance profilers.

RequestCpuTimeInterceptor interceptor will set the CPU execution time of each request in a Prometheus histogram using cpu profiler.

Both interceptors expect an instance of metricsService class as an argument.

import { MetricsService, RequestCpuTimeInterceptor, LoggingInterceptor } from '@multiversx/sdk-nestjs-monitoring';

async function bootstrap() {
  // AppModule imports MetricsModule
  const publicApp = await NestFactory.create(AppModule);
  const metricsService = publicApp.get<MetricsService>(MetricsService);

  const globalInterceptors = [];
  globalInterceptors.push(new RequestCpuTimeInterceptor(metricsService));
  globalInterceptors.push(new LoggingInterceptor(metricsService));


MetricsModule and MetricsService

MetricsModule is a Nestjs Module responsible for aggregating metrics data through MetricsService and exposing them to be consumed by Prometheus. MetricsService is extensible, you can define and aggregate your own metrics and expose them. By default it exposes a set of metrics created by the interceptors specified here. Most of the Multiversx packages expose metrics by default through this service. For example @multiversx/sdk-nestjs-redis automatically tracks the execution time of each redis query, overall redis health and much more, by leveraging the MetricsService.

How to instantiate the MetricsModule and expose metrics endpoints for Prometheus

In our example we will showcase how to expose response time and CPU time of HTTP requests. Make sure you have the interceptors in place as shown here. After the interceptors are in place, as requests comes through your application, the metrics are being populated into MetricsService class and we just have to expose the output of the .getMetrics() method on MetricsService through a controller.

import { Controller, Get } from '@nestjs/common';
import { MetricsService } from '@multiversx/sdk-nestjs-monitoring';

export class MetricsController {
    private readonly metricsService: MetricsService

  getMetrics(): string {
    return this.metricsService.getMetrics();

How to add custom metrics

Adding custom metrics is just a matter of creating another class which uses MetricsService.

We can create a new class called ApiMetricsService which will have a new custom metric heartbeatsHistogram.

import { Injectable } from '@nestjs/common';
import { MetricsService } from '@multiversx/sdk-nestjs-monitoring';
import { register, Histogram } from 'prom-client';

export class ApiMetricsService {
  private static heartbeatsHistogram: Histogram<string>;

  constructor(private readonly metricsService: MetricsService) {
    if (!ApiMetricsService.heartbeatsHistogram) {
      ApiMetricsService.heartbeatsHistogram = new Histogram({
        name: 'heartbeats',
        help: 'Heartbeats',
        labelNames: ['app'],
        buckets: [],

  async getMetrics(): Promise<string> {
    const baseMetrics = await this.metricsService.getMetrics();
    const currentMetrics = await register.metrics();

    return baseMetrics + '\n' + currentMetrics;

  setHeartbeatDuration(app: string, duration: number) {

The only change we have to do is that we need to instantiate this class and call .getMetrics() method on it to return to us both default and our new custom metrics.

The .setHeartbeatDuration() method will be used in our business logic whenever we want to add a new value to that histogram.

Package Sidebar


npm i @multiversx/sdk-nestjs-monitoring

Weekly Downloads






Unpacked Size

204 kB

Total Files


Last publish


  • dragosrebegea
  • arhtudormorar
  • radumojic
  • tanghel
  • cipriand
  • ccorcoveanu
  • valentinlup
  • andreibancioiu
  • titel
  • lucian.mincu