amplify-education
serverless-domain-manager
Create custom domain names that your lambda can deploy to with serverless. Allows for base path mapping when deploying and deletion of domain names.
About Amplify
Amplify builds innovative and compelling digital educational products that empower teachers and students across the country. We have a long history as the leading innovator in K-12 education - and have been described as the best tech company in education and the best education company in tech. While others try to shrink the learning experience into the technology, we use technology to expand what is possible in real classrooms with real students and teachers.
Learn more at https://www.amplify.com
Getting Started
Prerequisites
Make sure you have the following installed before starting:
The IAM role that is deploying the lambda might need the following permissions:
acm:ListCertificates *
acm:DescribeCertificate *
apigateway:AddCertificateToDomain /domainnames*
apigateway:RemoveCertificateFromDomain /domainnames*
apigateway:GET /domainnames*, /apis*, /restapis*
apigateway:DELETE /domainnames*, /apis*, /restapis*
apigateway:POST /domainnames*, /apis*, /restapis*
apigateway:PATCH /domainnames*, /apis*, /restapis*
cloudformation:GET *
cloudformation:ListStacks *
cloudformation:DescribeStacks *
cloudfront:UpdateDistribution *
route53:ListHostedZones *
route53:ChangeResourceRecordSets hostedzone/{HostedZoneId}
route53:GetHostedZone *
route53:ListResourceRecordSets *
iam:CreateServiceLinkedRole arn:aws:iam::${AWS::AccountId}: role/aws-service-role/ops.apigateway.amazonaws.com/AWSServiceRoleForAPIGateway
s3:ListBucket *
s3:GetObject *CloudFormation
Alternatively you can generate an least privileged IAM Managed Policy for deployment with this:
deployment policy cloudformation template
Installing
# From npm (recommended)
npm install serverless-domain-manager --save-devThen make the following edits to your serverless.yaml file:
Add the plugin.
plugins:
- serverless-domain-managerAdd the plugin configuration (example for serverless.foo.com/api). For a single domain and API type the following structure can be used.
custom:
customDomain:
domainName: serverless.foo.com
stage: ci
basePath: api
certificateName: '*.foo.com'
createRoute53Record: true
createRoute53IPv6Record: true
endpointType: REGIONAL
securityPolicy: tls_1_2
apiType: rest
autoDomain: falseMultiple API types mapped to different domains can also be supported with the following structure. The key is the API Gateway API type.
custom:
customDomain:
rest:
domainName: rest.serverless.foo.com
stage: ci
basePath: api
certificateName: '*.foo.com'
createRoute53Record: true
createRoute53IPv6Record: true
endpointType: REGIONAL
securityPolicy: tls_1_2
http:
domainName: http.serverless.foo.com
stage: ci
basePath: api
certificateName: '*.foo.com'
createRoute53Record: true
createRoute53IPv6Record: true
endpointType: REGIONAL
securityPolicy: tls_1_2
websocket:
domainName: ws.serverless.foo.com
stage: ci
basePath: api
certificateName: '*.foo.com'
createRoute53Record: true
createRoute53IPv6Record: true
endpointType: REGIONAL
securityPolicy: tls_1_2Or for multiple domains
custom:
customDomains:
- http:
domainName: http-api-${opt:RANDOM_STRING}.${env:TEST_DOMAIN}
basePath: ''
endpointType: REGIONAL
- http:
domainName: http-api-${opt:RANDOM_STRING}.${env:TEST_DOMAIN}.foo
basePath: ''
endpointType: REGIONALFor multi-region deployments, a route53Params structure can be used to support latency or weighted routing policies
custom:
customDomain:
domainName: serverless.foo.com
stage: ci
basePath: api
certificateName: '*.foo.com'
createRoute53Record: true
endpointType: REGIONAL
securityPolicy: tls_1_2
route53Params:
routingPolicy: latency| Parameter Name | Default Value | Description |
|---|---|---|
| domainName (Required) | | The domain name to be created in API Gateway and Route53 (if enabled) for this API. | |
| basePath | (none) |
The base path that will prepend all API endpoints. |
| stage | Value of --stage, or provider.stage (serverless will default to dev if unset) |
The stage to create the domain name for. This parameter allows you to specify a different stage for the domain name than the stage specified for the serverless deployment. |
| certificateName | Closest match | The name of a specific certificate from Certificate Manager to use with this API. If not specified, the closest match will be used (i.e. for a given domain name api.example.com, a certificate for api.example.com will take precedence over a *.example.com certificate). Note: Edge-optimized endpoints require that the certificate be located in us-east-1 to be used with the CloudFront distribution. |
| certificateArn | (none) |
The arn of a specific certificate from Certificate Manager to use with this API. |
| createRoute53Record | true |
Toggles whether or not the plugin will create A Alias and AAAA Alias records in Route53 mapping the domainName to the generated distribution domain name. If false, does not create a record. |
| createRoute53IPv6Record | true |
Toggles whether or not the plugin will create an AAAA Alias record in Route53 mapping the domainName to the generated distribution domain name. If false, does not create a record. |
| route53Profile | (none) |
Profile to use for accessing Route53 resources when Route53 records are in a different account |
| route53Region | (none) |
Region to send Route53 services requests to (only applicable if also using route53Profile option) |
| endpointType | EDGE |
Defines the endpoint type, accepts REGIONAL or EDGE. |
| apiType | rest | Defines the api type, accepts rest, http or websocket. |
| tlsTruststoreUri | undefined |
An Amazon S3 url that specifies the truststore for mutual TLS authentication, for example s3://bucket-name/key-name. The truststore can contain certificates from public or private certificate authorities. Be aware mutual TLS is only available for regional APIs. |
| tlsTruststoreVersion | undefined |
The version of the S3 object that contains your truststore. To specify a version, you must have versioning enabled for the S3 bucket. |
| hostedZoneId | | If hostedZoneId is set the route53 record set will be created in the matching zone, otherwise the hosted zone will be figured out from the domainName (hosted zone with matching domain). | |
| hostedZonePrivate | | If hostedZonePrivate is set to true then only private hosted zones will be used for route 53 records. If it is set to false then only public hosted zones will be used for route53 records. Setting this parameter is specially useful if you have multiple hosted zones with the same domain name (e.g. a public and a private one). If records need to be set in both private and public hosted zones use splitHorizonDns parameter. |
|
| splitHorizonDns | false |
When hostedZoneId and hostedZonePrivate are not set, setting this to true creates route53 records in both private and public hosted zones with matching domain. |
| enabled | true | Sometimes there are stages for which is not desired to have custom domain names. This flag allows the developer to disable the plugin for such cases. Accepts either boolean or string values and defaults to true for backwards compatibility. |
| securityPolicy | tls_1_2 | The security policy to apply to the custom domain name. Accepts tls_1_0 or tls_1_2 |
| allowPathMatching | false | When updating an existing api mapping this will match on the basePath instead of the API ID to find existing mappings for an update. |
| autoDomain | false |
Toggles whether or not the plugin will run create_domain/delete_domain as part of sls deploy/remove so that multiple commands are not required. |
| autoDomainWaitFor | 120 |
How long to wait for create_domain to finish before starting deployment if domain does not exist immediately. |
| route53Params | | A set of options to customize Route 53 record creation. If left empty, A and AAAA records with simple routing will be created. If createRoute53Record is false, anything passed here will be ignored. |
|
| route53Params: routingPolicy |
simple | Defines the Route 53 routing policy, accepts simple, latency or weighted. |
| route53Params: weight |
200 |
Sets the weight for weighted routing. Ignored for simple and latency routing. |
| route53Params: setIdentifier |
| A unique identifier for records in a set of Route 53 records with the same domain name. Only relevant for latency and weighted routing. Defaults to the regional endpoint if not provided. |
|
| route53Params: healthCheckId |
| An optional id for a Route 53 health check. If it is failing, Route 53 will stop routing to it. Only relevant for latency and weighted routing. If it is not provided, no health check will be associated with the record. |
|
| preserveExternalPathMappings | false |
When autoDomain is set to true, and a deployment is removed, setting this to true checks for additional API Gateway base path mappings before automatically deleting the domain, and avoids doing so if they exist. |
Running
To create the custom domain:
serverless create_domainTo deploy with the custom domain:
serverless deployTo remove the created custom domain:
serverless delete_domainHow it works
Creating the custom domain takes advantage of Amazon's Certificate Manager to assign a certificate to the given domain name. Based on already created certificate names, the plugin will search for the certificate that resembles the custom domain's name the most and assign the ARN to that domain name. The plugin then creates the proper A Alias and AAAA Alias records for the domain through Route 53. Once the domain name is set it takes up to 40 minutes before it is initialized. After the certificate is initialized, sls deploy will create the base path mapping and assign the lambda to the custom domain name through CloudFront. All resources are created independent of CloudFormation. However, deploying will also output the domain name and distribution domain name to the CloudFormation stack outputs under the keys DomainName and DistributionDomainName, respectively.
Behavior Change in Version 3
In version 3, we decided to create/update/delete all resources through the API. Previously, only the basepath mapping was managed through CloudFormation. We moved away from creating anything through the stack for two reasons.
1) It seemed cleaner to have all resources be created in the same fashion, rather than just having one created elsewhere. Since multiple CloudFormation stacks can't create the same custom domain, we decided to have everything be done through the API.
2) We ran into issues such as #57 where the CloudFormation wasn't always being applied.
However, we still add the domain name and distribution domain name to the CloudFormation outputs, preserving the functionality requested in #43 implemented in #47.
Running Tests
To run unit tests:
npm testTo run integration tests, set an environment variable TEST_DOMAIN to the domain you will be testing for (i.e. example.com if creating a domain for api.example.com).
And ROUTE53_PROFILE for creating route53 record in one AWS account and deploy in another. Then,
export TEST_DOMAIN=example.com
export ROUTE53_PROFILE=default
export TLS_TRUSTSTORE_URI=s3://bucket-name/key-name
export TLS_TRUSTSTORE_VERSION=default
npm run build
npm run integration-testAll tests should pass. All unit tests should pass before merging. Integration tests will take an extremely long time, as DNS records have to propogate for the resources created - therefore, integration tests will not be run on every commit.
If there is an error update the node_modules inside the serverless-domain-manager folder:
npm installWriting Integration Tests
Unit tests are found in test/unit-tests. Integration tests are found in test/integration-tests. Each folder in tests/integration-tests contains the serverless-domain-manager configuration being tested. To create a new integration test, create a new folder for the handler.js and serverless.yml with the same naming convention and update deploy.test.ts or create a separate one with the test.ts ending.
Changing API Types
AWS API Gateway has three different API types: REST, HTTP, and WebSocket. Special steps need to be taken when migrating from one api type to another. A common migration will be from a REST API to an HTTP API given the potential cost savings. Below are the steps required to change from REST to HTTP. A similar process can be applied for other API type migrations.
REST to HTTP
1) Confirm the Domain name is a Regional domain name. Edge domains are not supported by AWS for HTTP APIs. See this guide for migrating an edge-optimized custom domain name to regional.
2) Wait for all DNS changes to take effect/propagate and ensure all traffic is being routed to the regional domain name before proceeding.
3) Make sure you have setup new or modified existing routes to use httpApi event in your serverless.yml file.
4) Make the following changes to the customDomain properties in the serverless.yml confg:
```yaml
endpointType: REGIONAL
apiType: http
allowPathMatching: true # Only for one deploy
```5) Run sls deploy
6) Remove the allowPathMatching option, it should only be used once when migrating a base path from one API type to another.
NOTE: Always test this process in a lower level staging or development environment before performing it in production.
Known Issues
- (5/23/2017) CloudFormation does not support changing the base path from empty to something or vice a versa. You must run
sls removeto remove the base path mapping. - (1/17/2018) The
create_domaincommand provided by this plugin does not currently update an existing Custom Domain's configuration. Instead, it only supports updating the Route 53 record pointing to the Custom Domain. For example, one must delete and recreate a Custom Domain to migrate it from regional to edge or vice versa, or to modify the certificate. - (8/22/2018) Creating a custom domain creates a CloudFront Distribution behind the scenes for fronting your API Gateway. This CloudFront Distribution is managed by AWS and cannot be viewed/managed by you. This is not a bug, but a quirk of how the Custom Domain feature works in API Gateway.
- (2/12/2019) Users who upgraded from 2.x.x to version 3.0.4 (now unpublished) and then reverted back to 2.x.x will be unable to deploy because of a bug that will be fixed in 3.1.0. The workaround is to delete the basepath mapping manually, which will let them successfully revert back to 2.x.x.
- (1/20/2022) Using
route53Profileoption requires having hosted zone for the domain in this profile and ACM certificate in the main profile (where functions are deployed). - (2/13/2024) ACM certificate must exist in the
us-east-1for theEDGEendpoint type. https://docs.aws.amazon.com/apigateway/latest/developerguide/how-to-edge-optimized-custom-domain-name.html
Responsible Disclosure
If you have any security issue to report, contact project maintainers privately. You can reach us at github@amplify.com
Contributing
We welcome pull requests! For your pull request to be accepted smoothly, we suggest that you:
- For any sizable change, first open a GitHub issue to discuss your idea.
- Create a pull request. Explain why you want to make the change and what it’s for. We’ll try to answer any PR’s promptly.