serverless-components / website Goto Github PK

Currently, all the assets come without any max-age headers. This leads to a lower performance score on Google Page Insights.

Any possibility to add those with this component?

I keep getting this error:

Error: Your AWS ACM Certificate for the domain "domain.com" has an unsupported status of: "VALIDATION_TIMED_OUT".  Please remove it manually and deploy again.

My serverless.yml is like

name: website

website:
  component: '@serverless/website'
  inputs:
    code:
      src: build
      hook: yarn build
    domain: sub.domain.com
    region: eu-west-1

Maybe is there an issue assigning a subdomain to the website?

When I deploy using @serverless/website I am unable to override the Bucket Policy for the S3 bucket I am creating.

Instead of it being:

"Principal": {
    "AWS": "*"
}

I would like it to be configurable to achieve:

"Principal": "*"

Is this achievable?
I would say yes since I installed the NPM package locally, and I was able to find the method configureBucketForHosting inside utils.js, the Principal value is hardcoded.

Is there a way for me to create a pull request to @serverless/website?
I am not able to find the open source repo

Hey, I'm trying to use variable to declare my domain name, but for some reason I can't seem to get it to work. Here's a simple example that errors

custom:
  domain: example.com

inputs:
  domain: ${self:custom.domain}

Which returns the error:

Error: "${self:custom.domain}" is not a valid domain.

Are variables not available for config?

Thanks

Is there a way to change the directory name the template is dropped into? For example, when I type serverless init react-starter, it will always put it in the react-starter directory. If I want to drop the template into the myAwesomeWebsite directory instead, is there a way to modify the init command to do this or do I just have to rename the directory every time?

Sorry if this is answered somewhere else, I dug through the components documentation and there doesn't seem to be any other commands documented with the init command.

Please update to support CL options and ${property.variables}

I'm having problems deploying my site to different AWS accounts for staging purposes. I have two AWS accounts, one for DEV and one for PROD. My current approach to deploy to either one is by running sls deploy --stage <stage> --aws-account <account>. Each AWS account stored within .aws/credentials is pointing to the corresponding IAM user with admin access. It seems like the --aws-account flag is being ignored and the bucket is created in the account associated with the provider account linked within the Serverless dashboard.

I also tried adding a provider for each account for my app within the dashboard following https://www.serverless.com/framework/docs/guides/providers/. By doing this I'm able to deploy other components (database, api, permissions) to the right accounts but the website component keeps failing with an Access denied error.

So am I missing something? My goal is ultimately to wire this up to a CI/CD pipeline - what is the correct way to deploy to different AWS accounts/stages?

Hi!

I am hitting this error on app.serverless.com (ci/cd deploys):

Error: Failed running "src.hook": "npm run build" due to the following error: Command failed: npm run build

Using Create React App with TypeScript ("react": "^17.0.1" && "typescript": "^4.0.3",):

org: usernameorg
app: app
component: website
name: app
#
stage: ${env:AWS_STAGE}
#
# https://github.com/serverless/components#simplicity
#
inputs:
  name: app-${stage}
  region: ${env:AWS_REGION}
  src:
    src: ./src
    hook: npm run build
    dist: ./build

  "scripts": {
    ...
    "build": "react-scripts build",
    "deploy:development": "export AWS_STAGE='development' && export AWS_REGION='us-west-2' && sls deploy",
    "remove:development": "export AWS_STAGE='development' && export AWS_REGION='us-west-2' && sls remove",
    ...
  },

app.serverless.com (ci/cd deploys) error

Build step: serverless deploy --stage development --region us-west-2 --force --org usernameorg --app app --debug
Initializing...
Action: "deploy" - Stage: "development" - Org: "usernameorg" - App: "app" - Name: "app"
Deploying...
 Error: Failed running "src.hook": "npm run build" due to the following error: Command failed: npm run build
npm ERR! code ELIFECYCLE
npm ERR! errno 1
npm ERR! [email protected] build: `react-scripts build`
npm ERR! Exit status 1
npm ERR! 
npm ERR! Failed at the [email protected] build script.
npm ERR! This is probably not a problem with npm. There is likely additional logging output above.
npm ERR! A complete log of this run can be found in:
npm ERR!     /root/.npm/_logs/2021-01-20T00_31_11_903Z-debug.log
    at exec (/var/lang/lib/node_modules/serverless/node_modules/@serverless/platform-client/src/instance.js:200:13)
    at ChildProcess.exithandler (child_process.js:301:5)
    at ChildProcess.emit (events.js:198:13)
    at maybeClose (internal/child_process.js:982:16)
    at Process.ChildProcess._handle.onexit (internal/child_process.js:259:5)
6s › Serverless › Failed running "src.hook": "npm run build" due to the following error: Command failed: npm run build
npm ERR! code ELIFECYCLE
npm ERR! errno 1
npm ERR! [email protected] build: `react-scripts build`
npm ERR! Exit status 1
npm ERR! 
npm ERR! Failed at the [email protected] build script.
npm ERR! This is probably not a problem with npm. There is likely additional logging output above.
npm ERR! A complete log of this run can be found in:
npm ERR!     /root/.npm/_logs/2021-01-20T00_31_11_903Z-debug.log
 
  Documentation: https://github.com/serverless/components 
  Support: https://app.serverless.com/support 
  Slack: https://www.serverless.com/slack/ 
Build step failed: serverless deploy --stage development --region us-west-2 --force --org usernameorg --app app --debug
Deploy failed!
Error: Build step failed: serverless deploy --stage development --region us-west-2 --force --org usernameorg --app app --debug

following the example, I'm getting this error:

Domain <my domain> was not found in your AWS account. Please purchase it from Route53 first then try again.
    at getDomainHostedZoneId (/Users/<me>/.serverless/components/registry/npm/@serverless/[email protected]/node_modules/@serverless/domain/utils.js:105:11)

found this: https://www.npmjs.com/package/@serverless/domain

but there's no link to a repo

so, this component fails to create ssl certificates if the domain a tld that has . in it. For example: .co.nz

Here is what happens:

serverless.yml as per documentation/example

component: website
name: some-name
stage: dev

inputs:
  domain: subdomain.example.co.nz
  src:
    src: ./src
    hook: npm run build
    dist: ./dist

run: serverless deploy --debug

error:

...
Checking if a certificate for the co.nz domain exists
Certificate for the co.nz domain does not exist. Creating...
Validating the certificate for the co.nz domain.

 TypeError: Cannot read property 'Name' of undefined
    at ensureCertificate (/var/task/utils.js:392:51)
    at processTicksAndRejections (internal/process/task_queues.js:97:5)
    at async Website.deploy (/var/task/serverless.js:59:37)
    at async Runtime.module.exports [as handler] (/opt/nodejs/node_modules/@serverless/
core/handler.js:273:24)

When i look at my AWS Certificate Manager, it is trying to create a SSL certificate for .co.nz (and *.co.nz) rather than example.co.nz (and *.example.co.nz). In the above scenario, i have example.co.nz on my Route53 Hosted Zones.

I can easily create a certificate manually, is there a way to tell the component to skip ssl certificate creation and simply provide the certificate arn?

I'm trying to deploy a simple React app in ap-southeast-2 but having trouble accessing the s3 transfer acceleration endpoint.

Additionally, buckets used with Amazon S3 Transfer Acceleration can't have dots (.) in their names according to these docs https://docs.aws.amazon.com/AmazonS3/latest/dev/BucketRestrictions.html where by default, my bucket name has period because I am trying to declare a domain.

Is it possible to turn off S3 Transfer Acceleration? Not sure if others are having this issue.

Also

➜ sls deploy --debug   
Initializing...
Action: "deploy" - Stage: "dev" - App: "reactapp-sls-components" - Instance: "reactapp-sls-components"
Deploying...
...
Deploying Website
Configuring bucket for hosting
Uploading Website files
 UnknownEndpoint: Inaccessible host: `s3-accelerate.amazonaws.com'. This service may not be available in the `ap-southeast-2' region.
    at Request.ENOTFOUND_ERROR (/var/runtime/node_modules/aws-sdk/lib/event_listeners.js:495:46)
    at Request.callListeners (/var/runtime/node_modules/aws-sdk/lib/sequential_executor.js:106:20)
    at Request.emit (/var/runtime/node_modules/aws-sdk/lib/sequential_executor.js:78:10)
    at Request.emit (/var/runtime/node_modules/aws-sdk/lib/request.js:683:14)
    at ClientRequest.error (/var/runtime/node_modules/aws-sdk/lib/event_listeners.js:333:22)
    at ClientRequest.<anonymous> (/var/runtime/node_modules/aws-sdk/lib/http/node.js:96:19)
    at ClientRequest.emit (events.js:311:20)
    at ClientRequest.EventEmitter.emit (domain.js:482:12)
    at TLSSocket.socketErrorListener (_http_client.js:426:9)
    at TLSSocket.emit (events.js:311:20)

4s › reactapp-sls-components › Inaccessible host: `s3-accelerate.amazonaws.com'. This service may not be available in the `ap-southeast-2' region.

Issue: Getting
AccessDenied: User: arn:aws:sts::XXXXXXXXXXXXXX:assumed-role/XXXXXXXX is not authorized to perform: iam:CreateRole on resource: arn:aws:iam::XXXXXXXXXX:role/<RANDOM_NAME>
while deploying a ui, api and db as some where down the chain the framework is trying to create roles and it is failing.

I think it is ideal to provide an ability for specifying a role name that already exists in the account or provide an ability to create role and use that for deployment.

In certain organizations the role creation might have restriction.
Ex: Allow creation of new roles only if a standard permission boundary is attached to it. or other scenarios wrt to roles...

In those lines custom role creation will always run into issues.. So if an IAM role is provided, on the fly role creation should be skipped and provided role should be attached to the resources.

@eahefnawy @pmuens @medikoo Thoughts ?

The folders are not uploaded/not created on S3. The folder names are merged into the filenames as string (flattened), therefore static assets are not displayed on the created website, worst case with create-react-app that application will not even start because static chunk JavaScript files are created in the static folder.

When I upload the static folder via the AWS console manually into the same bucket then everything works fine.

Probably the problem is here: websiteBucket.upload.

Btw where can i find this @serverless/aws-s3 package which has the websiteBucket.upload function on github? May be I could do a PR.
Thanks

Trying to deploy a website that would be hosted on a.sub.domain.com, I have a route53 and certificate setup for sub.domain.com but due to company policy cannot have that on domain.com
While deploying it translates the domain naked in util.js resulting to domain.com which fails. It would be better to give a nakedDomain config on serverless.yml. Can you also suggest a way to overcome this problem?

Hello, I am new to the Serverless community. Serverless framework has been awesome for me. I migrated my API(s) successfully completely through Serverless framework tutorial. Now I am migrating my front-end code as well. As I am building my react front-end app using this website component, I stumbled upon how I can configure dev vs prod environment. To be precise, under my /src folder I have appconfig.js file where I have environment variables logic with process.env.NODE_ENV that tells the front-end which endpoint it should go (dev/prod). With Serverless website (Client-side rendering) and static hosting, is there any way I can point my front-end to different API endpoints depending on the stage or some other variables? If not, will this feature be added? Thank you for the help in advance!

Route 53 does not have IPv6 AAAA record after deploy. Would be better to have a way to configure (enable/disable) IPv6.

The idea here is to just keep the react-starter template up to date with the latest library versions and the current "preferred" way to develop react components, using functional components and hooks over class based components and state.

I have set things up according to the example with my own custom domain. However, the domain 301 redirects to the aws generated bucket url? Where the site is loading fine, but obviously i don't want the redirect. is this intended behaviour?

The name of the bucket used to store the static files is saved in the .serverless folder, which isn't supposed to be included in source control. Therefore, every time a new version of a website is deployed from a machine that doesn't contain that folder (CD/CI pipeline for instance), the component creates a new bucket where it uploads the website. This isn't ideal in some cases, especially since old buckets aren't automatically removed. Therefore, I believe it would make sense to let people pass the name of a concrete bucket to the component through an input parameter (same as region for instance).

I've got a fork of the component with that change made, and I'd be more than happy to submit a PR if you agree with the proposed solution.

Result from : sls deploy

Updating CloudFront distribution of ID XXXXXXXX.
Adding "xxxxx.com" certificate to CloudFront distribution
Adding domain "www.xxxxx.com" to CloudFront distribution
Adding domain "xxxxx.com" to CloudFront distribution

I dont own the CNAME for www.xxxxx.com so it throws a InvalidViewerCertificate

serverless.yml

component: website
name: landing-page4
stage: dev

inputs:
  domain: 'http://www.xxxxx.com'
  src:
    src: ./src
    hook: npm run build
    dist: ./dist

Why is it adding the www domain? I cant deploy to live as it stands as its forcing this options. Appreciate some help

Hello everyone !

I encountered an issue that I solved after a long suffering.

The issue

I deployed a nuxt project to AWS by using the website component in the full static nuxt generate mode.
I had incoherent results like if all the pages had only access to the payload that was generated for the "/" page.

I reproduced the issue here : https://d2xk172pvk6ejw.cloudfront.net/

To reproduce the bug :

• go to the cloudfront distribution
• go to Secondary page go back to Home page : you can see there are two different messages (an email for the homepage, a message with emojis for secondary)
• now refresh the browser while being on the Secondary page => you get the email as on the home page.

You can find the code here : https://github.com/oussj/nuxt-full-static-serverless-website-issue

After looking a bit more, I understood that it was linked to this issue : nuxt/nuxt#2506.

I understood it was linked to the fact that the cloudfront distribution is created with the s3 bucket url, and not the s3 website url. As you can see here in the source code

When I went to the AWS console and I manually changed the Origin Domain Name, everything was back to normal.

Suggestion

I suggest to always create the cloudfront distribution with the website url http://website-oecg92n.s3-website-us-east-1.amazonaws.com instead of http://website-oecg92n.s3.us-east-1.amazonaws.com

If this is not okay for some cloudfront reason (I'd be very interested to understand it), I'd suggest to add a boolean optional input static in serverless.yml

If static would be set to true, the cloudfront distribution would be created with the s3 website url and not the s3 bucket url.

I'd be happy to do the pull request!

When switching to a new AWS environment, we kept getting this error:

Error: Forbidden: Invalid credentials or this AWS S3 bucket name may already be taken

Credentials were correct. Bucket wasn't taken. If there is a way to return AWS error codes to the user that would be more than a little useful. Currently, one must recreate an S3 bucket and associate it to a CF CDN via the console to find the actual error: ACM:listCertificates

Extra Credit: Improve documentation to list precise policies needed to run.

I don't know if this is the right place for this - the error looks like it's in platform-client - but that is a private repo.

I get this intermittent error when I deploy from CI/CD environment:

Build step: serverless deploy --stage dev --region us-east-1 --force --org darbio --app tsz-next --debug
Initializing...
Action: "deploy" - Stage: "dev" - Org: "darbio" - App: "tsz-next" - Name: "frontend"
Deploying...
/var/lang/lib/node_modules/serverless/node_modules/@serverless/platform-client/src/Connection.js:114
            this.connection.ping();
                            ^
TypeError: Cannot read property 'ping' of null
    at Timeout.pingInterval.setInterval [as _onTimeout] (/var/lang/lib/node_modules/serverless/node_modules/@serverless/platform-client/src/Connection.js:114:29)
    at ontimeout (timers.js:436:11)
    at tryOnTimeout (timers.js:300:5)
    at listOnTimeout (timers.js:263:5)
    at Timer.processTimers (timers.js:223:10)
Build step failed: serverless deploy --stage dev --region us-east-1 --force --org darbio --app tsz-next --debug

I think there is a bug in platform-client when the connection times out.

Here is the output of serverless command :

this is the output:

website:
url: http://b8s9a8h-8wmi9k.s3-website-eu-central-1.amazonaws.com
env:

The returned url doesn't work/exists.

Thank you

Hi there, first off thanks for making such an easy to use framework!

I'm trying out this component for deploying a React SPA with S3 & CloudFront. However I'm noticing that build artifacts from previous builds remain in my bucket - is there an option to remove these on a fresh deploy? (Similar to the default behaviour of serverless-finch)

Best
Zac

Hello!

Most of the suggestion is in the title. When I audited with Lighthouse my static website deployed with the serverless website component, I found out that I had a lot to win by gzipping my assets.

Gzipping is not enabled by default when creating the distribution as I could see it here

I suggest to set this value to true by default instead of false

Otherwise, if for some reason the default behaviour should not change, I suggest adding it as an optional input in the serverless.yml file

I'd be happy to do the pull request.

It would be better if we create a CloudFront Origin Access Identity and grant permission to get Object from S3 bucket instead of using AWS principal.

On the most recent Serverless version, had this S3 bucket name as an input:

bucketName: welcome-dev.remember.us

The correct s3 bucket is created correctly:

http://welcome-dev.remember.us.s3-website-us-west-2.amazonaws.com

I also mapped a subdomain, and it points to a cloudfront dist that was also created.

The problem is that the cloudfront dist has the origin as welcome-dev.s3.amazonaws.com which is wrong. It's not using the whole bucket name welcome-dev.remember.us.

When trying to set a wildcard domain you get an error saying the domain isn't a valid domain.

ex in yaml:

domain: "*.myapp.com"

result:

41s » Serverless » "*.myapp.com" is not a valid domain.

From the code the website url is http://${bucketOutputs.name}.s3-website-${inputs.region}.amazonaws.com. Shouldn't it be http://${bucketOutputs.name}.s3-website.${inputs.region}.amazonaws.com? ('.' instead of '-')

What's the best way to change api key or api domain uri based on a different stage in a single page application?

If this were a lambda, I would use environment variables in the serverless yml file, but since these are static files I'm not sure the best way to do this.

In the company I work for we are deploying two websites using this component (configured for CloudFront) at the same time. In some occasions, we get a Rate exceeded error, which breaks the deployment, and we can't really tell by the logs where it is coming from.

This is a screenshot of the error:

Do you have any idea what could be the cause of the issue? I've digged into the code and I can't really tell what AWS could be returning the error...

HI,

I'm working with this serverless component, so I'm trying to create an artifact with this command

serverless package --package .artifact

but I'm receiving several issues and it seems it doesn't work.

any chance to support that?

Hi,
I have two AWS accounts a master account for production and a sub account for Development. In production the deployment works fine, but when I try to deploy in the development account it fails with the following error:

Serverless › [RRSet with DNS name _3638471d4d0b2191ad3718504521cbe6.mydomain.com. is not permitted in zone dev.mydomain.com.]

In my main account (prod): the domain is like that : backoffice.mydomain.com (works fine) <---- subdomain of "mydomain.com"
in my sub account (dev): it's like that : backoffice.dev.mydomain.com (failed) <---- subdomain of "dev.mydomain.com"

For more detail, In my sub account I have only one hosted zone "dev.mydomain.com" , I have delegated this domain to sub account from the master account.

serverless.yml

Hey,

I'm playing around this component, and it seems to be working great 👍 .

One thing I cannot get my head around is different environments if I hook this up to some kind of CI process.

For example, if I wanted to support dev, staging, and production

How would you go about doing this?

My thought was to change the bucket name that the code gets pushed too, but it looks like I cannot do this with cli options like I could in the standard serverless.yml setup?

My vue-starter project no longer gives public read access to the access bucket when it's deployed, I have to manually make everything in the bucket public after every deploy.

Here are the steps I took to get into this state.

1. Create a new project from the vue-starter template
2. Run "sls" from the command line to deploy the website
3. Open the serverless.yml and add a bucketName and domain that match.
4. Change the app: vue-starter and the name: vue-starter to a new value
5. Run "sls" again. Observe error that the bucket already exists for a different app and you should remove it first and redeploy.
6. Stash the changes to the yml, and run "sls remove"
7. Unstash your changes again and run "sls"
8. Log into s3 and Observe that all the files in the bucket no longer have Read permissions for the Everyone group.
9. Manually use the "Make Public" action in s3 to make everything available to the public.
10. Run "sls" again to redeploy
11. Refresh your website and see the 403 errors again, confirming in s3 that the files are no longer public again.

Following the tutorial for a www -> root domain redirect (e.g www.example.com -> example.com) https://serverless-stack.com/chapters/setup-www-domain-redirect.html, it directs you to set up a new cloudfront distribution for the www subdomain.

This means the www subdomain can no longer be an alternate domain name (CNAME) record for this cloudfront distribution or this error occurs on deploy:

Error: The domain "www.example.com" is already in use by another website or CloudFront Distribution.
    at updateCloudFrontDistribution (/var/task/utils.js:688:13)
    at processTicksAndRejections (internal/process/task_queues.js:97:5)

This is caused by

I suggest an option should be added to not include this www alternate domain name (CNAME) for the cloudfront distribution as it makes this style of www redirect impossible.

I've read the documents and it showed to me and attribute called env. Where are the env.js file after serverless deploy?

I'm using component: website vue-starter

Hi I have a static website project with a folder structure as follows:

Gruntfile.js		compiled/		package-lock.json	serverless.yml		yarn.lock
README.md		metadata.json		package.json		src/

I just want to upload the compiled/ files to my website/bucket but it uploads everything including the node_modules.

Here is my serverless.yml file:

website:
  component: "@serverless/website"
  inputs:
    src:
      src: ./src          # the input source folder
      hook: yarn && yarn build # the hook you want to run
      dist: ./compiled        # the output dist folder
    bucketName: ${cdn.bucketName}
    # domain: www.example.com

cdn:
  component: "@serverless/cdn"

After remove the project by using sls remove and remove all app from my Serverless account, then when I tried to deploy again I got the message

10s › Serverless › ENOENT: no such file or directory, scandir '/tmp/amazvd' 

Hi,
It would be awesome to allow origin configuration for the cloudfront distribution. It is a common use case when you want to place your API under xxxxxx.yy/api. That way you could use the output from deploying an API (i.e. through express component).

Right now, in serverless service I am able to do this through this cloudformation configuration:

            DomainName:
              Fn::Join:
                - ''
                - 
                  - Ref: ApiGatewayRestApi
                  - '.execute-api.${self:provider.region}.amazonaws.com'
            OriginPath: '/${opt:stage}'

I have tried to create my serverless blog with custom domain sls-blog.myowndomain.xyz

    []
    bucketName: 'sls-blog.myowndomain.xyz' 
    domain: sls-blog.myowndomain.xyz
    []

But it ends up with the error as below when I access my own domain

<Code>NoSuchBucket</Code>
<Message>The specified bucket does not exist</Message>
<BucketName>sls-blog</BucketName>

When I navigate to the Cloudfront Console, I recognize that the CloudFront distribution is trying access the origin as sls-blog.s3.amazonaws.com. Which I really have no chance to own it.

I tried to change the origin to the s3 bucket website domain and It seems to be working now.

Is Serverless components mistaking on create cloudfront distribution ?
I have check the code and find nothing seem to be wrong.

Hello sls!

I have been working with sls to deploy api's for the last few weeks without many problems.

Today I decided to start looking at serverless front-end, i found the serverless learn front end tutorial and it seems out of date. The components no longer exist, so I came to this github.

I tried following the instructions here in the README, and it still fails when trying to serverless deploy.

I tried creating a brand new empty project and running...

serverless init vue-starter
(this finishes fine)

cd vue-starter

serverless deploy
Error: {"errorMessage":"Authentication failed. - Please contact support and provide this identifier to reference this issue - ZVQH6BXLX8QJ"}

I tried using the same access key and secret that works for my api deploys and that didn't work, I also tried creating a new set of keys for the user (which already has admin permissions) and tried those, but it seems no matter what i do, it doesn't seem to work with them in the .env file, and it also isn't using the keys that are in the ~/.aws/credentials file

Given the following website serverless.yml the bucket is named a random system assigned GUID name regardless of if the bucketName property is there or not.
`
name: my.domain.com
stage: local

website:
component: '@serverless/website'
inputs:
code:
src: client/dist
hook: domain=my.domain.com npm run clean:build
region: us-east-1
bucketName: local.my.domain.com-specialbucketname123
env:
API_URL: https://my.domain.com
STAGE: local

domain: my.domain.com

`
Tried:

• With/without quotes
• deleting and re-creating buckets via S3 console
• bucket with desired name was able to be created via console as well.

The .env file is non-standard and requires scripting to switch out or tokenize the .env for different AWS environments. Please update to support 'profile' in yml that references ~/.aws/credentials entries

I've set up a Vue project and have the template set up according to the example. It creates a public bucket but does not upload anything to the bucket. I have the code root folder pointed to the base directory and the src directory pointed to the dist folder that is created after "npm run build".

for a project I'm working on we discovered it'd be desirable to override the indexDocument and errorDocument - and would be great to be able to specify the desired override in the serverless.yml/serverless.js configuration, eg, something like:

# serverless.yml

name: my-website
stage: dev

myWebsite:
  component: "@serverless/website"
  inputs:
    code:
      root: ./ # The root folder of your website project.  Defaults to current working directory
      src: ./src # The folder to be uploaded containing your built artifact
      hook: npm run build # A hook to build/test/do anything to your code before uploading
    region: us-east-1 # The AWS region to deploy your website into
    bucketName: myBucket # (Optional) The Bucket name where `src` files/folder will be upload. 
                         # If not provided, it will create random bucket name and upload `src` files
    env: # Environment variables to include in a 'env.js' file with your uploaded code.
      API_URL: https://api.com

    ### NEW PROPERTIES HERE
    indexDocument: index.html # (Optional) the default page for your website
    errorDocument: index.html # (Optional) the error page for your website.
    ### END NEW PROPERTIES


    # You can specify a custom domain name for your website.
    # You must have a public hosted zone available for this domain in AWS Route53.
    # This is done automatically for you if you've purchased the domain via AWS Route53.
    domain: www.example.com

Any chance I could submit a patch to this project to enable this functionality?

When my config contains a domain entry with a dash:

domain: www-test.mydomainname.com

the deploy fails with

InvalidChangeBatch: [The request contains an invalid set of changes for a resource record set 'A www-test.mydomainname.com.']

but when the config has no dash

domain: test.mydomainname.com

it works fine.

After installing and trying to run npm run build I got the error:
Duplicate plugin/preset detected.

I was able to resolve this by removing the plugins section from package.json

since the preset nano-react-app already includes those plugins, they don't need to be declared again in package.json heres a related stackoverflow explaining