Deploy NestJS with Fastify on Google Cloud Run

Repository: WhatDaFox/nestjs-fastify-cloud-run-poc

Configure Google Cloud

To be able to build and deploy, you will need a Google Cloud project, with a billing account set up, as well as
the Google Cloud CLI installed.

Then you will need to create a configuration for your project:

$ gcloud config configurations create cloud-run
$ gcloud auth login # and follow the steps
$ gcloud config set project YOUR_PROJECT_ID

Create the project

For this proof of concept, I will only use the default NestJS application, that contains a single endpoint /
returning Hello world!:

$ npm i -g @nestjs/cli
$ nest new cloud-run

Install the Fastify driver:

$ npm i --save @nestjs/platform-fastify

We need to update main.ts to make use of the Fastify driver. Also, Cloud Run will decide the port of our application, so we have to update the main.ts file to reference
the PORT environment variable, like so:

import { NestFactory } from '@nestjs/core';
import { FastifyAdapter, NestFastifyApplication } from '@nestjs/platform-fastify';
import { AppModule } from './app.module';

async function bootstrap() {
  const app = await NestFactory.create<NestFastifyApplication>(
    AppModule,
    new FastifyAdapter({ logger: true }),
  );
  app.enableCors();
  await app.listen(parseInt(process.env.PORT) || 3000, '0.0.0.0');
}

bootstrap();

Now we are ready to create the Dockerfile.

Create the Dockerfile

We need to containerize our application to be able to run on Cloud Run. Create a Dockerfile at the root of your project
and copy/paste the following:

For better performance, I decided to build the app before hand and run the start:prod command.

# Use the official lightweight Node.js 12 image.
# https://hub.docker.com/_/node
FROM node:12-alpine

# Create and change to the app directory.
WORKDIR /usr/src/app

# Copy application dependency manifests to the container image.
# A wildcard is used to ensure both package.json AND package-lock.json are copied.
# Copying this separately prevents re-running npm install on every code change.
COPY package*.json ./

# Install dependencies
RUN npm install

# Copy local code to the container image.
COPY . ./

# Build the application
RUN npm run build

# Run the web service on container startup.
CMD [ "npm", "run", "start:prod" ]

Build & Deploy

Now, we can use Cloud Build to build our docker image. Cloud Build will automatically detect our Dockerfile, build,
and push our image in Google Container Registry:

$ gcloud builds submit --tag gcr.io/YOUR_PROJECT/helloworld

Once that's done, we can run the following command to deploy our new revision to Cloud Run:

$ gcloud run deploy --image gcr.io/YOUR_PROJECT/helloworld --platform managed

Benchmark

When testing, I ran a small (to avoid crazy costs) benchmark with Apache Benchmark.

Here is the command I ran:

$ ab -n 1000 -c 80 https://cloud-run-url/

Here are the results:

This is ApacheBench, Version 2.3 <$Revision: 1843412 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking cloud-run-url (be patient)
Completed 100 requests
Completed 200 requests
Completed 300 requests
Completed 400 requests
Completed 500 requests
Completed 600 requests
Completed 700 requests
Completed 800 requests
Completed 900 requests
Completed 1000 requests
Finished 1000 requests


Server Software:        Google
Server Hostname:        cloud-run-url
Server Port:            443
SSL/TLS Protocol:       TLSv1.2,ECDHE-RSA-CHACHA20-POLY1305,2048,256
Server Temp Key:        ECDH X25519 253 bits
TLS Server Name:        cloud-run-url

Document Path:          /
Document Length:        12 bytes

Concurrency Level:      80
Time taken for tests:   9.300 seconds
Complete requests:      1000
Failed requests:        0
Total transferred:      437004 bytes
HTML transferred:       12000 bytes
Requests per second:    107.53 [#/sec] (mean)
Time per request:       743.985 [ms] (mean)
Time per request:       9.300 [ms] (mean, across all concurrent requests)
Transfer rate:          45.89 [Kbytes/sec] received

Connection Times (ms)
              min  mean[+/-sd] median   max
Connect:       79  461 213.7    430    2633
Processing:    37  208 118.9    200     506
Waiting:       22  163 105.3    139     501
Total:        129  669 220.7    626    2739

Percentage of the requests served within a certain time (ms)
  50%    626
  66%    702
  75%    768
  80%    772
  90%    862
  95%   1161
  98%   1371
  99%   1576
 100%   2739 (longest request)

Conclusion

Compared to my previous experiment with the default installation of NestJS, I didn't observe any improvement in response time.

Blog