Meteor Boilerplate

A fast and highly customisable Meteor.js boiler plate app. The UI is handled by Material Design and the performance is achieved by selective tree-shaking and judicious use of dynamic imports, to keep the client bundle as small as reasonably possible.

>>> View Demo <<<

Front end: React ✌️
UI: MUI (Material Design) 🎨
Bundle size: <250kB gZipped 🐭
Styling: JSS 💅
Security: Users, roles and groups 🔐
Email: Just hook-up your favourite SMTP provider 📧
Developer friendly: you can use the start:ui cmd to just load the UI, for a snappy developer experience with Hot Module Reloading. It generally takes bout 50ms for a change to be built and loaded in the browser, which is about 3s quicker than if you were using the Meteor bundler (i.e. with npm start) 🤓
λ functions: Execute intensive functions in the cloud, making sure that your node server is always free to fulfil client requests 😅

For some background information on the performance enhancements in this app, check out this blog post.

Start Guide

Note that the meteor app lives in the /app directory. The root directory contains some general setup, such as linting, and leaves space for including other items such as end-to-end tests, a react native app and serverless functions.

Installing dependencies

Use npm to install the dependencies. Note that there are a few package.json files in the project.

npm i
cd serverless
npm i
cd ../app
meteor npm i

Starting the app

To run the full-blown app - front-end, back-end and MongoDB database, from the app directory run:

npm start

If you just want to work on the UI, then, from the app directory, you can run:

npm run start:ui

λ functions

This projects uses Zeit Now to deploy and execute lambda functions.

Meteor runs a Node server; it is single threaded process and we want to make sure that big computations don't grind the server to a halt. Hence the λ functions. Any big computations should be sent to a λ function. The beauty of this is that we can spin up as many λ functions at once and not need to even worry about scaling our app.

Dependencies for the serverless functions are defined in serverless/package.json. Make sure they are installed by running

cd serverless
npm install

Running λ functions locally

You must first run the Now development environment:

From the app directory, you can use the shortcut

npm run now:dev

Or do it manually, from the serverless directory. Your choice.

now dev --local-config=development.now.json --port 8080

After which you are able to call those functions from your code. e.g.

        HTTP.call(
          'POST',
          `${Meteor.settings.public.serverless_url}/lambdas/inc.js`,
          { data: { amount: 5 } },
          (err, res) => {
            if (err) {
              snacks.handleMethodError(err);
            } else {
              snacks.setMessage(res.content);
            }
          },
        );

Using secrets with λ functions

Any secret keys that need to be available to your labmda functions can be accessed through the node environment, e.g. process.env.MONGO_URL.

These keys are set in the relevant .now.json files. For the dev environment (i.e. dev.now.json) you can just paste your secrets directly into the relevant env section of the file. However, for live deployments, you need to create your secret with Now and then reference that secret from your .now.json file. For example, if your production Google API key is "123xzy" then run the following to create a secret with Now:

now secret add production-google-api-key 123xyz

and then in your production.now.json file add a reference to it:

{
  "env": {
    "GOOGLE_API_KEY": "@production-google-api-key"
  }
}

Deploying λ functions

Create an account with Zeit and install the NOW desktop client or CLI. Then, from the serverless directory deploy your lambda functions by running:

now --local-config=production.now.json

Teams

Note that if you are a member of one or more teams, then you should run the following before deploying, to make sure that you deploy to the correct team

now switch <team-slug|username>

ninjaPixel/meteor-boilerplate