v4.9.0
SIGN UP
LOGIN
FIDDLEFORUMSSTATUS
SUPPORT

This guide shows you how to deploy a Nuxt.js application on Layer0. If you run into any issues please consult the Troubleshooting section.

This Nuxt.js example app uses server-side rendering and prefetching to provide lightening-fast transitions between pages.

This Nuxt.js example app uses ISG (Incremental Static Generation) to provide lightening-fast transitions between pages.

This framework has a connector developed for Layer0. See Connectors for more information.

If you have not already done so, install the Layer0 CLI

npm i -g @layer0/cli

If you don't already have a nuxt.js application, you can create one using:

npm create nuxt-app my-nuxt-app

Nuxt's create module will ask you a series of questions to configure your app. Make sure you answer as follows:

  • For Choose rendering mode select Universal (SSR)
  • Your answers to the other questions should not matter for the purposes of this guide.

To prepare your Nuxt.js application for Layer0:

  1. In the existing nuxt.config.js configuration, add "@layer0/nuxt/module" to buildModules:
// nuxt.config.js

module.exports = {
  ...
  buildModules: [['@layer0/nuxt/module', { layer0SourceMaps: true }]],
  ...
}

Options:

  • layer0SourceMaps: true|false: when true, the serverless build includes sourcemap files which make debugging easier when tailing the server logs in the Layer0 Developer Console. It also increases the serverless bundle size, which may push your deployments over the 50MB (compressed) limit.
  1. Run 0 init to configure your project for Layer0.
0 init

The 0 init command will automatically add all the required dependencies and files to your project. These include:

  • The @layer0/core package
  • The @layer0/nuxt package
  • The @layer0/vue package
  • layer0.config.js - Contains various configuration options for Layer0.
  • routes.js - A default routes file that sends all requests to nuxt.js. You can update this file to add caching or proxy some URLs to a different origin as described later in this guide.
  • sw/service-worker.js - A service worker that provides static asset and API prefetching.

This command will also update your package.json with the following changes:

  • Moves all packages in dependencies to devDependencies except those listed in the modules property of nuxt.config.js.
  • Adds @nuxt/core to dependencies
  • Adds several scripts to run the available 0 commands

As an example, here's the original package.json from Nuxt's create step:

{
  "name": "my-nuxt-app",
  "version": "1.0.0",
  "description": "My remarkable Nuxt.js project",
  "author": "Techy Ted",
  "private": true,
  "scripts": {
    "dev": "nuxt",
    "build": "nuxt build",
    "start": "nuxt start",
    "generate": "nuxt generate"
  },
  "dependencies": {
    "@layer0/cli": "^2.0.0",
    "@layer0/core": "^2.0.0",
    "@layer0/nuxt": "^2.0.0",
    "nuxt": "^2.0.0"
  },
  "devDependencies": {}
}

And here is the package.json after modifications by 0 init:

{
  "name": "my-nuxt-app",
  "version": "1.0.0",
  "description": "My remarkable Nuxt.js project",
  "author": "Techy Ted",
  "private": true,
  "scripts": {
    "dev": "0 run",
    "build": "0 build",
    "start": "0 run",
    "prod": "0 run --production",
    "generate": "nuxt generate"
  },
  "dependencies": {
    "@nuxt/core": "^2.14.2"
  },
  "devDependencies": {
    "@layer0/cli": "^2.0.0",
    "@layer0/core": "^2.0.0",
    "@layer0/nuxt": "^2.0.0",
    "@layer0/vue": "^2.0.0",
    "dotenv": "^8.2.0",
    "nuxt": "^2.0.0",
    "serverless": "^1.64.0",
    "serverless-dotenv-plugin": "^2.3.2",
    "serverless-offline": "^5.14.1"
  }
}

Run the Nuxt.js app with the command:

npm run 0:dev

Load the site: http://127.0.0.1:3000

Nuxt does not bundle packages listed in the modules property of nuxt.config.js when building your app for production. This can lead to an increased bundle size and slow down server-side rendering. Most Nuxt modules can be moved to buildModules. We recommend the following to maximize performance of server-side rendering in the cloud:

  • Move all entries from modules to buildModules in nuxt.config.js
  • Move all corresponding packages from dependencies to devDependencies in package.json
  • Run yarn install or npm install to update your lock file.

Doing so will exclude these modules from your production deployment and keep the bundle size as small as possible.

The next few sections of this guide explain how Layer0 interacts with Nuxt's routing, which is important if you are migrating an existing application. If you just created a new nuxt app, you can jump to Running Locally and come back to these sections later. Layer0 supports Nuxt.js's built-in routing scheme. The default routes.js file created by 0 init sends all requests to Nuxt.js via a fallback route:

// This file was automatically added by 0 deploy.
// You should commit this file to source control.
const { Router } = require('@layer0/core/router')
const { nuxtRoutes, renderNuxtPage } = require('@layer0/nuxt')

module.exports = new Router().use(nuxtRoutes)

In the code above, nuxtRoutes adds all Nuxt.js routes based on the /pages directory. It's also compatible with extending Nuxt's router via the router config in nuxt.config.js, for example:

// nuxt.config.js
export default {
  // ... more config ...
  router: {
    // For example, we can extend the nuxt router to accept /products in addition to /p.
    // The nuxtRoutes middleware automatically picks this up and adds it to the Layer0 router
    extendRoutes(routes, resolve) {
      routes.push({
        path: '/products/:id?',
        component: resolve(__dirname, 'pages/p/_id.vue'),
      })
    },
  },
  // ... more config ...
}

You can add additional routes before and after nuxtRoutes, for example to send some URLs to an alternate backend. This is useful for gradually replacing an existing site with a new Nuxt.js app.

A popular use case is to fallback to a legacy site for any route that your Nuxt.js app isn't configured to handle:

new Router().use(nuxtRoutes).fallback(({ proxy }) => proxy('legacy'))

To configure the legacy backend, use layer0.config.js:

module.exports = {
  backends: {
    legacy: {
      domainOrIp: process.env.LEGACY_BACKEND_DOMAIN || 'legacy.my-site.com',
      hostHeader: process.env.LEGACY_BACKEND_HOST_HEADER || 'legacy.my-site.com',
    },
  },
}

Using environment variables here allows you to configure different legacy domains for each Layer0 environment.

The easiest way to add edge caching to your nuxt.js app is to add caching routes before the middleware. For example, imagine you have /pages/c/_categoryId.js:

new Router()
  .get('/pages/c/:categoryId', ({ cache }) => {
    cache({
      browser: {
        maxAgeSeconds: 0,
        serviceWorkerSeconds: 60 * 60 * 24,
      },
      edge: {
        maxAgeSeconds: 60 * 60 * 24,
        staleWhileRevalidateSeconds: 60 * 60,
      },
    })
  })
  .use(nuxtMiddleware)

The @layer0/nuxt/module builds a service worker that enables prefetching using Layer0 and injects it into your app's browser code. The service worker is based on Google's Workbox library. The entry point for the service worker source code is sw/service-worker.js. If your app has an existing service worker that uses workbox, you can copy its contents into sw/service-worker.js and simply add the following to your service worker:

import { Prefetcher } from '@layer0/prefetch/sw'
new Prefetcher().route()

The above allows you to prefetch pages from Layer0's edge cache to greatly improve browsing speed. To prefetch a page, add the Prefetch component from @layer0/vue to any router-link or nuxt-link element:

<template>
  <ul v-for="product in products">
    <li>
      <Prefetch v-bind:url="'/api/' + product.url">
        <nuxt-link v-bind:to="product.url">
          <img v-bind:src="product.thumbnail" />
        </nuxt-link>
      </Prefetch>
    </li>
  </ul>
</template>
<script>
  import { Prefetch } from '@layer0/vue'
  export default {
    components: {
      Prefetch,
    },
  }
</script>

The Prefetch component fetches data for the linked page from Layer0's edge cache based on the url property and adds it to the service worker's cache when the link becomes visible in the viewport. When the user taps on the link, the page transition will be instantaneous because the browser won't need to fetch data from the network.

Layer0 supports fully and partially static sites using Nuxt generate. To deploy a static Nuxt site on Layer0, simply set target: 'static' in nuxt.config.js and run 0 deploy. This will run nuxt build and nuxt generate to generate a static version of your site.

By default, requests for any pages that are not statically rendered at build time will fall back to server side rendering. If you use the Layer0 router to cache pages that are not statically rendered, the first user who attempts to access the page will see the fallback HTML page generated by Nuxt (200.html by default). Layer0 will render and cache the HTML in the background so that subsequent visits result in a full HTML response. This behavior is similar to Next.js incremental static rendering (ISG). Here is an example route that adds caching for a partially static page:

import { Router } from '@layer0/core/router'
import { nuxtRoutes } from '@layer0/nuxt'

export default new Router()
  .get('/products/:id', ({ cache }) => {
    cache({
      edge: {
        // Requests for product pages that are not statically generated will fall back to SSR.
        // The first user will see the 200.html loading page generated by Nuxt.
        // Layer0 will render full HTML response in the background and cache it for one hour at the edge.
        // All future requests to the page will result in the full HTML response.
        maxAgeSeconds: 60 * 60 * 24,
        staleWhileRevalidateSeconds: 60 * 60, // continue to serve stale responses from the edge cache while refreshing via SSR in the background
      },
    })
  })
  .use(nuxtRoutes)

If you set the fallback property in the generate config to true, Nuxt.js will generate a 404.html page that will be served whenever the URL does not match a static page. Layer0 will send a 404 http status for these URLs. Note that if you set the fallback property to a string, Nuxt will generate a fallback page with that name, and Layer0 will serve it with a 200 http status when the URL does not match a statically generated page.

Nuxt requires that certain resources are included in a build and deploy to have access to them. As such, at times this will require additional configuration. To include additional resources for server side rendering, API calls, etc., use the includeFiles option in your layer0.config.js file. Read more

In this example, we would have an api folder that we want to include all items from.

includeFiles: {
  'api/**/*': true,
},

In addition, if includeNodeModules does not copy over the necessary package that may be needed in production, it can be included via this key as well. For instance,

includeFiles: {
  'node_modules/some_package/**/*': true,
}

The Nuxt team provides a renderer called Nitro which optimizes your application for serverless deployment and greatly minimizes the size of your server application bundle. If you're running into the size limitation for serverless bundles (50MB), you might try adding Nitro to your app. As of June 2021 Nitro is still not production ready, so use at your own risk.

Layer0 provides a connector specifically for Nuxt apps that use nitro called @layer0/nuxt-nitro.

To add Nitro to your app, make the following changes:

  1. Install nitro and the connector as dev dependencies:
npm install -D @nuxt/nitro @layer0/nuxt-nitro`
  1. Ensure buildModules in nuxt.config.js contains the following:
  buildModules: [
    '@nuxt/nitro/compat',
    '@layer0/nuxt-nitro/module', // If you have previously added @layer0/nuxt/module you can remove it.
    // ...others...
  ],
  1. Add the following to nuxt.config.js:
  publicRuntimeConfig: {
    nitroVersion: require('@nuxt/nitro/package.json').version,
  },

  1. If your nuxt.config.js has a target property, remove it.

  2. If you've previously added @layer0/nuxt as a dependency, you can remove it.

To test your app locally, run:

0 build && 0 run

You can do a production build of your app and test it locally using:

0 build && 0 run --production

Setting --production runs your app exactly as it will be uploaded to the Layer0 cloud using serverless-offline.

Deploying requires an account on Layer0. Sign up here for free. Once you have an account, you can deploy to Layer0 by running the following in the root folder of your project:

0 deploy

See deploying for more information.

The following section describes common gotchas and their workarounds.

This may be because you have a custom server framework (such as Express). Please make sure you selected None when asked to choose Choose custom server framework during the creation of your nuxt app.

If you get a command not found error such as:

$ 0 init
- bash: 0: command not found

Make sure you installed the Layer0 CLI

npm i -g @layer0/cli

If you previously installed the Layer0 CLI, make sure your version is current.

Check npm for the latest released version of the CLI:

$ npm show @layer0/cli version
1.16.2

Compare the latest release against the version currently installed on your system:

$ 0 --version
1.16.2

If your version is out of date you can update it by running

npm update -g @layer0/cli

As the error states, there is an upper limit on how big a package can be when deployed to our serverless infrastructure. Some common strategies for solving:

  • You may need to move some dependencies as described here. Only dependencies are copied up to the lambda.
  • Make sure you are using imports in a smart way. A common example is changing: import { get } from lodash to import get from lodash/get to avoid unnecessary bloat in your modules

You can view what is included in your package under .layer0/lambda/ after a build, and running du -h -d 1 on the directories in a shell will output the size of each directory and help you identify where space savings can be found, ie du -h -d 1 .layer0/lambda/.nuxt

Edit this guide on GitHub