Hi, my use case is to show a field to all people connected via HTML. If it updates, it needs to reflect on everyone's page. Is this a good starting point, or should I be looking at GraphQL?

Little puzzled why you don't have an HTML interface in your sample here.

Thanks to this fix to use a supported nodejs version, it should be possible to use this sample application again. However, this requires a manual release. Could someone do this please?

Client samples mentioned in blog are no longer valid.

1. In the blog we have $request.body.action as route selection expression while latest code in this repo has $request.body.message as route selection expression.
2. Correct route for message sending is sendMessage, not sendmessage as mentioned in blog post.

Originally posted by @armujahid in #6 (comment)

Steps to reproduce

1. Install the template in this repo
2. Create a Custom Domain in the API Gateway
3. Attach created websocket api into to API mappings in the domain
4. Setup the CNAME entry in your domain DNS
5. RUN following command

   wscat -c wss://<your.custom.domain>
   > {"action": "sendmessage", "data":"hello"}
   

   This message does not return 'hello' back
   The Cloudwatch logs for the SendMessage logs show that the message is received by the server and there is no error in sending the reply back

I have been attempting to port this chat-app to a aws-cdk application, and feel like i'm close, but not quite there yet.

This is the response i'm getting
`[ec2-user@ip-172-31-x-x websocket]$ wscat -c wss://xxxxxxx.execute-api.ap-southeast-2.amazonaws.com/prod
Connected (press CTRL+C to quit)

{"action":"sendmessage","data":"hello world"}
< {"message": "Internal server error", "connectionId":"VNRUSeldSwMCIRQ=", "requestId":"VNRXVFJOywMF3-Q="}`

I know that the message is arriving at the lambda associated with the sendmessage lambda. I'm able to to print it and see it in the lambda logs. I'm getting a client error when it trys to post_connection.

ERROR] EndpointConnectionError: Could not connect to the endpoint URL: "https://execute-api.ap-southeast-2.amazonaws.com/@connections/VNmugd9KSwMCFwQ%3D"

I have attempted to reduce my lambda down to a simple as seems sensible. Just to try and diganose the issue, i also gave the lambda function full administrative permissions, but that made no difference either.

Heres the labmda code.

import boto3
import os
from botocore.exceptions import ClientError
import json
def lambda_handler(event, context):
    message = 'one two three'.encode('utf-8')
    body = event['body']
    print(body)
    connectionId = event['requestContext']['connectionId']
    print(connectionId)
    api_client = boto3.client('apigatewaymanagementapi')
    api_client.post_to_connection(
            Data = message,
            ConnectionId = connectionId
    )
    return {    
            'statusCode': 200, 
            'body': 'Message Sent' 
    }

I have deployed the app to cloudformation, and when I try to connect to the wss endpoint, I get

Unexpected server response: 500

When I had the app installed from the automatic AWS deployment, it worked fine, but not when deployed via aws cloudformation

I have followed the instructions and used the AWS CLI to deploy the application - everything works fine - however, if you add a new route everything appears to have been created but API Gateway returns a forbidden message.

The only solution is to delete the stack and then re-run via the AWS CLI - the new route then works fine.

This may well be a Cloudformation issue or maybe someone can point me in the right direction. To create a new route I simply did the following:

• copy the sendmessage folder as sendmessagedupe and update the name in the package.json
• create a new route and integration in the resources section of the template with name and properties set appropriately
• add the route to the deployment -> dependson list
• create a new function and permission in the resources section of the template with name and properties set appropriately

(mirroring all the sections required by the original send message route and function)

As I say, deleting and recreating the stack then works so perhaps a AWS CLI/Cloudformation issue but not a workable solution as the stack is then created with a new url.

Hello I'm beginner of sam.
I want to run the server in my pc, and communicate with other device through socket.
I just run a 'sam local start-lambda', but I can't find the path for socket connection.

Connect route has $connect endpoint. Disconnect has $disconnect. What is the route for $customAuthorizer ?

I directly deployed this project through aws, and deployed it, I can connect using wscat -c url, but when I try to send any message I get a response (with actual strings for connectionid and request id):
{"message": "Forbidden", "connectionId":"some_string", "requestId":"some_string"}

Is there a step that needs to be applied outside of the deployment that is not documented? (I did definitely deploy the api and stage it)

Reported on twitter:

"It’s reporting that “it is in a rollback_complete state and cannot be updated”. This is happening even in a region that I haven’t attempted a deployment."

by using a IAM user with correct permissions, I was able to deploy in a different region that my default one. "famous works on my machine"

I'm attempting to test this repository local with sam local start-lambda --region us-east1, which spins up at 127.0.0.1:3001. However, any attempt to connect to to this path wtih wss:// or ws:// returns "Bad Request" within the SAM output.

It would be great if testing documentation could be added to the repository.

SendMessage lamda function is throwing exception if I send this message:
{"message":"sendMessage", "data":"hello world"}

I deployed it directly using serverless repo and haven't changed a single line of code.

Cloudwatch logs:

Complete logs:

START RequestId: 041b19b4-8d3f-4247-8d4c-16ba493b7b0f Version: $LATEST
Unable to import module 'app': Error
    at Function.Module._load (module.js:474:25)
    at Module.require (module.js:596:17)
    at require (internal/module.js:11:18)
    at Object.<anonymous> (/var/task/node_modules/aws-sdk/clients/apigatewaymanagementapi.js:1:63)
    at Module._compile (module.js:652:30)
    at Object.Module._extensions..js (module.js:663:10)
    at Module.load (module.js:565:32)
    at tryModuleLoad (module.js:505:12)
    at Function.Module._load (module.js:497:3)
END RequestId: 041b19b4-8d3f-4247-8d4c-16ba493b7b0f
REPORT RequestId: 041b19b4-8d3f-4247-8d4c-16ba493b7b0f	Duration: 1.85 ms	Billed Duration: 100 ms 	Memory Size: 256 MB	Max Memory Used: 84 MB	

This error is related to require('aws-sdk/clients/apigatewaymanagementapi');
"aws-sdk": "^2.404.0" is already present in dependencies so I don't know why this is happening.

There is section of code where the 'stale' connection is detected.
This is the code snippet. I am not using the Promise.all that u guys are using if it does affect in anyway. Just calling await normally. And the value of apiRes here is {}, without erroring out Is there any work around to this? Because now sometimes my messages goes into unknown darkness :(

      try {
        var apiRes = await apigwManagementApi.postToConnection({
          ConnectionId: connectionId,
          Data: JSON.stringify(sockBody.data)
        }).promise();
        console.log('apiRes',apiRes);
      } catch (e) {
        console.log('apigwManagementApi.postToConnection:catch:',e);
      }

When I built this sample, the API Gateway trigger integration shows up in Lambda console:

When building my own app, everything works just fine, but the trigger does not show up in Lambda console:

Again, I should point out that my API is working just fine, it's just odd and makes me feel like I might be missing something important.

My template looks pretty much the same (so does the API Gateway route overview in the console):

...

ConnectRoute:
    Type: AWS::ApiGatewayV2::Route
    Properties:
      ApiId: !Ref RedeChatWSApi
      RouteKey: $connect
      AuthorizationType: NONE
      OperationName: ConnectRoute
      Target: !Join
        - "/"
        - - "integrations"
          - !Ref ConnectInteg
  ConnectInteg:
    Type: AWS::ApiGatewayV2::Integration
    Properties:
      ApiId: !Ref RedeChatWSApi
      Description: Connect integration.
      IntegrationType: AWS_PROXY
      IntegrationUri:
        Fn::Sub:
          arn:aws:apigateway:${AWS::Region}:lambda:path/2015-03-31/functions/${OnConnectFunction.Arn}/invocations

...

Outputs:
  OnConnectFunctionArn:
    Description: OnConnect function ARN.
    Value: !GetAtt OnConnectFunction.Arn

Thanks for the help. 😀

When calling the sendmessage Lambda function, the following line:
await apigwManagementApi.postToConnection({ ConnectionId: connectionId, Data: postData }).promise();

returns an error:
No method found matching route test0/@connections/PxWP8cfSDoECFYw= for http method POST.

The connection ID is that of the caller, so should be able to mirror the message back to the caller.

Is this just a permissions issue, maybe the account calling the function doesn't have permission to post the the connection endpoint.

Hi,

thanks for the sample, which helped us a lot to understand how to use WebSockets with Amazon API Gateway. We were able to create a new Socket API for our solution based on the samples. We wanted to integrate it created via Amazon CloudFront into our existing solution (according to the AWS Documentation WebSockets should work on CloudFront: https://goo.gl/4AMRPb).

Currently we do have a CloudFront distribution (incl. CNAME with e.g. mydomain,com) with multiple origins

• an S3 Origin where we do have the frontend (frontendOrigin)
• a custom origin for the API (apiOrigin)
• an additional S3 Origin, where we also store frontend pictures (picturesOrigin)

For these origins we created several Cache Behaviours:

• The Default (*) behaviour is pointing to frontendOrigin
• A Behaviour with Path Pattern /api/* which is pointing to apiOrigin
• A Behaviour with Path Pattern /pictures/* which is pointing to picturesOrigin

Now we wanted to integrate our new Socket API created via API Gateway to CloudFront and created a new Origin called socketOrigin with a Behaviour with Pattern /api/socket which is pointing to socketOrigin (with the correct precedence).

After redeploying the distribution we were not able to open a connection to the socket, because of a ** HTTP 403** Status code. I found a way to get more details by using a curl:

curl --include \
     --no-buffer \
     --header "Connection: Upgrade" \
     --header "Upgrade: websocket" \
     --header "Host: <cloudfrontid>.cloudfront.net" \
     --header "Origin: https://<cloudfrontid>.cloudfront.net/api/socket" \
     --header "Sec-WebSocket-Key: SGVsbG8sIHdvcmxkIQ==" \
     --header "Sec-WebSocket-Version: 13" \
     https://<cloudfrontid>.cloudfront.net/api/socket

We got the following response back by CloudFront:

HTTP/1.1 403 Forbidden
Content-Type: application/json; charset=UTF-8
Content-Length: 75
Connection: close
Date: Thu, 10 Jan 2019 09:57:07 GMT
X-Cache: Error from cloudfront
Via: 1.1 blahblahblah.cloudfront.net (CloudFront)
X-Amz-Cf-Id: YxowUpxLm6t3VEV1Foue36kmdrOilhVlxqHN0S59dy-7cca681LiHw==

We thought, that it might have something to do with the path pattern, as we don't specify any wildcard, so we switched the path pattern to /api/socket/* , but this didn't help. We checked the Viewer Protocol Policy (HTTP & HTTPS) and Origin Protocol Policy (Match Viewer), but this also didn't help to solve the 403.

So, here is the question, if Amazon API Gateway & Amazon CloudFront is working with WebSocket APIs and if yes, how it must be configured to get to work. Can you provide any samples? If it is not working yet, we would like to open a feature request to support the above mentioned configuration.

Kind Regards,
Marcell

Why are you doing the following?

// Add ApiGatewayManagementApi to the AWS namespace
require('aws-sdk/clients/apigatewaymanagementapi');

So I simply deployed this code with no changes whatsoever. I can get a connection with no issues but when calling sendMessage I get an internal error. Looking in Cloudwatch, I see this error:

Unable to import module 'app': Error
at Module._compile (module.js:652:30)
at Object.Module._extensions..js (module.js:663:10)
at Module.load (module.js:565:32)
at tryModuleLoad (module.js:505:12)
at Function.Module._load (module.js:497:3)

This app has been modified in GitHub, but those updates haven't been published to SAR. Deploying the app via SAR is a much simpler experience than using GitHub and SAM CLI. SAR has some features that launched since this app was released that should help you keep this app in sync.

SAM CLI added a sam publish command that allows you to specify the SAR metadata of the app in the template itself. This will allow you to follow best practices like including the GitHub URL for the app and sourceCodeUrls, for example. Then you can run sam publish.

Opening an issue to add SAR metadata and use either sam publish or the new CodePipeline app to make sure the SAR app stays in sync with changes made to this GitHub repo.

• sam publish docs: https://docs.aws.amazon.com/serverless-application-model/latest/developerguide/sam-cli-command-reference-sam-publish.html
• SAR metadata docs: https://docs.aws.amazon.com/serverless-application-model/latest/developerguide/serverless-sam-template-publishing-applications-metadata-properties.html
• Example app metadata: https://github.com/jlhood/github-codebuild-logs/blob/master/template.yml#L7-L24
• CodePipeline tutorial for automatic publishing to SAR: https://docs.aws.amazon.com/codepipeline/latest/userguide/tutorials-serverlessrepo-auto-publish.html

Is there some way to serve Web socket locally to have sane development iterations?

I don't think it's supported by https://docs.aws.amazon.com/serverless-application-model/latest/developerguide/serverless-sam-cli-using-start-api.html

template.yaml references the code packages for the functions as camel case, but the actual directories are not camel case.

You can either rename the directories or change the template to fix this.

Can we enrich this sample by adding client code. So we can see how client will receive messages

While running the Lambda function sendmessage on AWS west-2; got the following error:

const AWS = require('aws-sdk');
// Add ApiGatewayManagementApi to the AWS namespace
require('aws-sdk/clients/apigatewaymanagementapi');

{
"errorMessage": "Cannot find module 'aws-sdk/clients/apigatewaymanagementapi'",
"errorType": "Error",
"stackTrace": [
"Function.Module._load (module.js:474:25)",
"Module.require (module.js:596:17)",
"require (internal/module.js:11:18)",
"Object. (/var/task/index.js:7:1)",
"Module._compile (module.js:652:30)",
"Object.Module._extensions..js (module.js:663:10)",
"Module.load (module.js:565:32)",
"tryModuleLoad (module.js:505:12)",
"Function.Module._load (module.js:497:3)"
]
Is ApiGatewayManagementApi not part of SDK version in Oregon?

When I do:

webSocket.send(
  JSON.stringify({
    action: 'sendmessage'
  })
);

I'm gonna get an event back with the following:

{
  "message": "Internal server error",
  "connectionId": <hidden>,
  "requestId": <hidden>
}

The deployment fails with:

"Resource handler returned message: "The runtime parameter of nodejs12.x is no longer supported for creating or updating AWS Lambda functions. We recommend you use the new runtime (nodejs18.x) while creating or updating functions. (Service: Lambda, Status Code: 400, Request ID: 0cfaea73-519e-4c9e-a57d-f983a2e2c252)" (RequestToken: 5b6cfae8-94b7-745c-c31d-f7029c02e83d, HandlerErrorCode: InvalidRequest)"

I have an app which uses redis to hold socketIds. Everytime when user leaves the site, $disconnect function is called in lambda and user is being removed from cache and something is logged to CloudWatch where I can check if function was called.

Unfortunately when I turn off the internet connection on client (browser/PWA), the $disconnect lambda is not triggered :(

how could this be possible?

I'm encountering a problem when sometimes my client connects to websockets API and sometimes not. I can't see anything from cloudwatch blog. Could that be an issue of the websockets gateway? Unfortunately I couldn't find anything about the error message.

Hello I have some issue to correctly setup the AWS::ApiGatewayV2::Authorizer for my API.

According to the documentation I create this:

Resources:
  DevWebSocket:
    Type: 'AWS::ApiGatewayV2::Api'
    Properties:
      Name: TL-Dev-WebSocket-API
      ProtocolType: WEBSOCKET
      RouteSelectionExpression: $request.body.action
  DevAuthorizerLambda:
    Type: 'AWS::Serverless::Function'
    Properties:
      CodeUri: WebSockets/Authorizer
      Role: 'arn:aws:iam::************:role/LambdaDynamoDB'
      Environment:
        Variables:
          STAGE: Dev
  DevAuthorizerLambdaPermission:
    Type: 'AWS::Lambda::Permission'
    Properties:
      Action: 'lambda:invokeFunction'
      Principal: apigateway.amazonaws.com
      FunctionName:
        Ref: DevAuthorizerLambda
      SourceArn:
        'Fn::Sub':
          - >-
            arn:aws:execute-api:${AWS::Region}:${AWS::AccountId}:${__ApiId__}/${__Stage__}/$connect
          - __Stage__: '*'
            __ApiId__:
              Ref: DevWebSocket
  DevWebSocketAuthorizer:
    Type: 'AWS::ApiGatewayV2::Authorizer'
    Properties:
      Name: DevAuthorizer
      ApiId:
        Ref: DevWebSocket
      AuthorizerType: REQUEST
      AuthorizerUri:
        'Fn::Sub': >-
          arn:aws:apigateway:${AWS::Region}:lambda:path/2015-03-31/functions/${DevAuthorizerLambda.Arn}/invocations
      IdentitySource:
        - route.request.querystring.token
  DevWebSocketDeployment:
    Type: 'AWS::ApiGatewayV2::Deployment'
    Properties:
      ApiId:
        Ref: DevWebSocket
    DependsOn:
      - WebSocketPart1 # ref routes to avoid error 'need a least one route to create this'
  DevWebSocketStage:
    Type: 'AWS::ApiGatewayV2::Stage'
    Properties:
      StageName: Dev
      Description: Dev
      DeploymentId:
        Ref: DevWebSocketDeployment
      ApiId:
        Ref: DevWebSocket
  [...]

But currently I get all the time Unauthorized and my lambda for authorization is not trigger at all (no cloudwacth logs).

Thank you in advance for your help.

Is there a way to run this locally? I get this error:

$simple-websockets-chat-app git:(master) sam local start-api               

Error: not enough values to unpack (expected 2, got 1)
Traceback:
  File "click/core.py", line 1055, in main
  File "click/core.py", line 1657, in invoke
  File "click/core.py", line 1657, in invoke
  File "click/core.py", line 1404, in invoke
  File "click/core.py", line 760, in invoke
  File "click/decorators.py", line 84, in new_func
  File "click/core.py", line 760, in invoke
  File "samcli/lib/telemetry/metric.py", line 184, in wrapped
  File "samcli/lib/telemetry/metric.py", line 149, in wrapped
  File "samcli/lib/utils/version_checker.py", line 42, in wrapped
  File "samcli/cli/main.py", line 92, in wrapper
  File "samcli/commands/local/start_api/cli.py", line 97, in cli
  File "samcli/commands/local/start_api/cli.py", line 193, in do_cli
  File "samcli/commands/local/lib/local_api_service.py", line 37, in __init__
  File "samcli/lib/providers/api_provider.py", line 37, in __init__
  File "samcli/lib/providers/api_provider.py", line 64, in _extract_api
  File "samcli/lib/providers/cfn_api_provider.py", line 88, in extract_resources
  File "samcli/lib/providers/cfn_api_provider.py", line 442, in _extract_cfn_gateway_v2_route
  File "samcli/lib/providers/cfn_api_provider.py", line 618, in _parse_route_key

Do Cloudformation templates currently support automatically creating API Gateway for WebSockets type and NOT REST type?

How do we close a specific WebSocket API Gateway client connection from the server side?

As reported from twitter feedback:

Should prob update this part of the code in the sample and blog post – the forEach loop is synchronous, so execution will continue past it without waiting for the callbacks. Def easier to use async/await here!

I see that the code in this repo has been updated to use node 10 and then node 12, but the link in this article (https://aws.amazon.com/blogs/compute/announcing-websocket-apis-in-amazon-api-gateway/) deploys node 8 - and fails.

As per the title, I have created a super simple script to connect to the wss:// endpoint provided, using socket.io

const socket = io('wss://<app_id>.execute-api.<region>.amazonaws.com/<stage_name>');
socket.emit('sendmessage', 'hi');

And I get the following error in the console:

Cross-Origin Request Blocked: The Same Origin Policy disallows reading the remote resource at https://<app_id>.execute-api..amazonaws.com/socket.io/?EIO=4&transport=polling&t=NZM3Y0O. (Reason: CORS header 'Access-Control-Allow-Origin' missing).

How can I add CORs support to the API Gateway Websocket? I cannot find any information on this.

When you change the ARN of a function or an integration, this template does not make a new deployment.

Try changing the logical name of the Function for one of the handlers, along with the references to it (to use the new logical name of the function), and repackage and deploy the SAM package.

No new deployment will be triggered, because the Deployment resource, even with the explicit DependsOn, will not make a new deployment. The old function (with the older logical name in the template) will be deleted, and the API will continue to try to access the old function's ARN.

There does not seem to be any way to force an in-place replacement safely of the Deployment resource. I have needed to change the logical name of the Deployment resource to be something like:

Deployment to work around this. It caused a pretty serious outage on our end when using this SAM pattern without that. Beware to folks using this template if you're doing anything that causes a full-replacement on any of the Lambda resources or API Gateway Integrations.