Deploy Gulp Yeoman webapp to Amazon S3

Recently I wrote a recipe for how to deploy a Yeoman Gulp Webapp to Amazon S3. Now I'd like to give more configuration on deploying to S3. Specifically, optimization notes.


After you follow the recipe and add gulp deploy task, I highly recommend you to revision your assets. The above recipe puts a long expiry date which means that once the assets have been downloaded to client's browser, they will stay there for a long time. Which is actually good, unless you don't make any changes.

There's a recipe for that: Asset revisioning. Just follow it, straigtforward.

This recipe demonstrates how to set up simple static asset revisioning (aka revving) for CSS and JS by appending content hash to their filenames unicorn.cssunicorn-098f6bcd.css.

Compression (gzip)

Compression should not be confused with minification. Gzipping can save up to 50% of the served files size. Which also means that it will save your bandwith (and hosting if you pay for) billing.

I tried to do it on my own project with gulp-gzip, but I had a hard time with Content-Types and Content-Encoding. Filename references can be solved with gulp-rev-replace. Which is also described in asset revisioning recipe.

The easiest solution I found was to use gulp-awspublish which already has a built-in support for it. We are already using it for deployment. Just add the following line and that's it:

gulp.task('deploy', ['build'], function () {
  // create a new publisher
  var publisher = $.awspublish.create({
    key: '...',
    secret: '...',
    bucket: '...'

  // define custom headers
  var headers = {
    'Cache-Control': 'max-age=315360000, no-transform, public'

  return gulp.src('dist/**/*.*')
+   .pipe($.awspublish.gzip({ ext: '' }))

It will gzip all your files, keeping their original filenames and set the correct Content-Encoding on Amazon S3.


comments powered by Disqus