- Create a REST API out of a json-server compatible JSON-File
- Swagger UI integrated (Swagger spec automatically generated)
- GrapqhiQL integrated (Graphql schema automatically generated)
- Additional JSON File validations at startup
- Readonly or Read/Write Mode (file stored in S3 Bucket)
- Deployment in AWS:
- Own CLI to test locally and immediately deploy it to the cloud
- Deployed in AWS cloud within Minutes by a single command
- Almost zero costs (First million requests for Lambda are free)
- Less maintenance as the deployed solution runs serverless
- Security:
- Secured with https by default.
- Optional: Use a generated API Key
- Customization:
- This solution written in Typescript can be easily extended for additional enhanced scenarios
- adding user authentication (Here`s an example)
- own custom domain
- additional routes etc.
- This solution written in Typescript can be easily extended for additional enhanced scenarios
npm i -g json-serverless
-
create a jsonserver-file sample e.g. db.json
{ "posts": [ { "id": 1, "title": "json-server", "author": "typicode" }, { "id": 2, "title": "test", "author": "yourAuthor" } ], "comments": [ { "id": 1, "body": "some comment", "postId": 1 } ], "profile": { "name": "typicode" } }
-
execute command
jsonsls run db.json
-
Verify that you have a AWS account and set appropriate credentials
-
execute command
jsonsls create-stack db.json {optional: STAGE}
- a stack template folder will be created that contains the deployable serverless framework solution. You can use the serverless cli in this stack template folder.
-
When the deployment was successful you can see following output
Service Information service: serverless-json-server stage: dev region: eu-central-1 stack: serverless-json-server-dev api keys: serverless-json-server.dev: {API-KEY} endpoints: ANY - https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/ <== {ENDPOINTURL} ANY - https://xxxxxxx.eu-central-1.amazonaws.com/dev/{proxy+} functions: app: serverless-json-server-dev-app layers: None Serverless: Removing old service artifacts from S3...
Open the {ENDPOINTURL}: https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/ that you received as output
Features | Relative Path | Sample with Endpoint |
---|---|---|
Swagger UI | /ui | https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/ui |
Swagger Specification | /api-spec | https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/api-spec |
GraphiQL | /graphql | https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/graphql |
API Routes | /api/{routes} | https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/api/posts |
MIND: If you have set enableApiKeyAuth to true => SwaggerUI )
- replace the url with the url provided by serverless (see above)
- replace the {API-KEY} with the key you get from serverless (see above)
- replace {route} at the end of the url e.g. with posts (default value)
Default Schema:
Default route is posts: (see db.json)
curl -H "Content-Type: application/json" https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/api/posts
# or another route given in db.json file
curl -H "Content-Type: application/json" https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/api/{route}
# with enableApiKeyAuth=true
curl -H "x-api-key: {API-KEY}" -H "Content-Type: application/json" https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/api/{route}
What`s my {route} ? -> see json-server documentation
Please have a look at this example to see how you can add own middleware and authentication.
https://github.com/pharindoko/jsonsls-vue-cognito-demo
-
update local db.json file in root directory with new values
-
re-deploy the stack via serverless framework
jsonsls update-stack
-
delete db.json file in S3 Bucket
-
Make a GET request against the root url https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/api
curl -H "Content-Type: application/json" https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/api
# with enableApiKeyAuth=true
curl -H "x-api-key: {API-KEY}" -H "Content-Type: application/json" https://xxxxxx.execute-api.eu-central-1.amazonaws.com/dev/api/{route}
=> With the next request a new db.json file will be created in the S3 Bucket
Attribute | Description | Type | Default |
---|---|---|---|
readOnly | Make API readonly - all API - write operations are forbidden (http 403)) | string | false |
enableApiKeyAuth | Make your routes private by using an additional ApiKey | boolean | false |
enableJSONValidation | validate JSON file at start | boolean | true |
enableSwagger | enable or disable Swagger (and related Graphql) features | boolean | true |
PRs are welcome!
Please have a look into the makefile to get the understanding how this construct is built. There are components managed: (under /packages)
-
cli
- built with oclif
- main purpose is to enease the usage of the tool
- creates a stack folder out of the template package
-
server
- the core component of the solution
- has json server implemented under the hood
- components will be injected from outside into the library (storageadapter, swagger)
- can be customized and used without the other parts (could be deployed with docker)
- library can be used standalone as well (see example)
-
template
-
a serverless framework template
-
can be used standalone without the cli
serverless create --template-url https://github.com/pharindoko/json-serverless/tree/master/packages/template npm i npm run build sls deploy sls offline
-
make install
Lerna will be used to manage the monorepo`s dependencies.
lerna bootstrap
make start-server
the json file will be loaded directly from your local filesystem. No AWS access is needed.
This part is using the serverless offline and is close to how the solution behaves in the cloud.
make start-template
the json file will be loaded directly from your local filesystem. No AWS access is needed.
you should see a direct output of all local endpoints.
make start-cli
the json file will be loaded directly from your local filesystem. No AWS access is needed.
The apiKey is set in AWS API Gateway. This means all requests (even the standard route) need to use the API-KEY.
If you want to see the Swagger UI you need to add a plugin e.g. ModHeader to Chrome and add the needed headers:
- Content-Type: application/json
- x-api-key: {provided by sls info in the output after deployment}
Ensure you have credentials for AWS set.
sls info
sls remove
Check Cloudwatch Logs in AWS - the issue should be describe there. Log has the same name as the stack that has been created.