anyone know if it’s possible in CDK to reference a Lambda layer that’s already deployed (outside of the infra code).. especially one in another AWS account

just leaving this here before I actually go to bed.. I’m having trouble granting a Lambda func permissions to two buckets:

transcodeFunction.attachPermissions([
    [props!.uploadBucket, 'grantRead'],
    [props!.storageBucket, 'grantWrite'],
  ]);

when I deploy, I get this:

TypeError: construct[methodName] is not a function

how do I test Lambda functions using

yarn start

if they’re not part of an API? 🤔 say I have 5 functions that are part of a step function, and I want to just test one of those in isolation. in Serverless Framework, I could just use

sls invoke local -f

to do this.

How do I get the current stage value inside the lambda function? i.e local, stage, prod and etc..

Hi, I am deploying my stack to production. However, viewing the cloudwatch logs, I get the following issue:

[ERROR] Runtime.ImportModuleError: Unable to import module 'index': No module named 'src' Traceback (most recent call last):

2021-08-12T15:06:09.638+01:00

My function definition is as follows:

const skillsSyncFunction = new Function(this, 'SkillsSyncFunction', {
      srcPath: "src/lambdas/skills_sync",
      handler: 'index.handler',
      timeout: 30,
      environment: env,
      permissions: ["ssm", skillsTable, bulkAssignOrchestratorTopic],
      description: "Blah Blah"
    });

Folder structure:

MyApp/src/lambdas/skills_sync/index.py

Everything works locally in the

Debug_Stack

I should add

morning, you fine folks! quick/easy question about monorepos + yarn workspaces. say I have 4 services in

src/services

, and 2 of them need the same npm dependency. should I run

yarn add somepkg

at the root level, or just do

yarn workspace service1 add somepkg && yarn workspace service2 add somepkg

? in other words, are root level dependencies intended to be shared among workspaces/services.. or should it be reserved only for things pertaining to the SST stack code?

Hey all, getting a bit of an error running the

start

command. Was working fine up until this morning. I start up and then I get a websocket connection error that just infinitely repeats. I don't really think I changed very much that would have affected this. Look familiar to anyone, or anyone know where I might be able to start looking? https://gist.github.com/mittonface/28080899524db2978e567e7aaefa4995

Hello, does anyone successfully ran the tests on serverless-stack/lerna-yarn-starter ?

Having some issues with the websocket api not entirely sure what I am doing wrong. In my stack I have:

new WebSocketApi(this, 'WebSocketApi', {
	routes: {
		$connect: 'src/functions/actions/websocket/connect.handler'
	}
})

And the lambda code is:

export const handler = (event: unknown) => {
	console.log(event)
	return {
		statusCode: 200,
		body: 'test'
	}
}

It logs the event just fine, but every websocket just immediately disconnects.

It looks like currently you cannot set the

routeResponseSelectionExpression

value for routes unless I am overlooking it, but it makes responding to websocket messages a lot easier. Without having that set you have to use the apigateway management api to respond whereas if I set it to

$default

whatever I put in the response body gets sent to the connection for me.

how do we avoid cyclic references in CDK/SST? for example, I set up Cognito in my “core” stack. later on, in my “media-processor” stack which depends on core, I need to add more permissions to authorized Cognito users:

props!.cognitoAuth.attachPermissionsForAuthUsers([
      [uploadProcessorFunc, 'grantInvoke'],
    ]);

This results in a cyclic dependency error.

Error: 'dev-microservices-core' depends on 'dev-microservices-media-processor' (dev-microservices-core -> dev-microservices-media-processor/UploadProcFunc/Resource.Arn). Adding this dependency (dev-microservices-media-processor -> dev-microservices-core/UploadBucket/Bucket/Resource.Arn) would create a cyclic reference.

is it safe to remove

"declaration": true,

from tsconfig? it was causing me some headaches

does

.build/

always contain the most recent set of

.js

and

.js.map

files after a build or deploy? in other words, is that a reliable place to have sentry-cli grab source & source maps from to upload to Sentry?

What is the workflow when multiple developer are working with SST. Do each developer need separate AWS account, or they can use same AWS account with different IAM user.

Hi all, Is it expected behaviour when running

npx sst remove

to get the following message:

dev-sst-integration-layer-my-stack failed: dev-sst-integration-layer-UsersTable already exists

I need to manually delete the dynamodb tables from AWS before running

npx sst remove

to ensure resource deletion.

potentially dumb question: do I have to run

yarn deploy

after stopping the

yarn start

process? any changes I’ve made during the debug process are already live anyway.. functions, at least. or does

yarn deploy

unhook the stubs & other magic needed to run stuff locally?

am I correct in thinking that

stack.setDefaultFunctionProps()

is the same thing as

app.setDefaultFunctionProps()

, except it only applies to functions created within that stack?

hey TypeScript savvy folks: how can I make it so both the infrastructure/stack code and my Lambda code stop yelling at me for using

process.env.<some var>

? I always have to use the non-null assertion. I’d rather just declare everything so I know I won’t make any mistakes.. e.g.

SENTRY_DSN: string;

just wanna bounce something off other devs (working solo on this stuff sucks, I miss water cooler days) 😄 say you have one SST project with a bunch of stacks. I’m thinking, does it make zero sense for Lambda functions to invoke other Lambda functions in the same or other stacks using

lambda.invokeFunction

? seems like it’s a better idea to just

import

the function and call it directly. I can’t think of any reasons off the top of my head why this is a bad idea.. 🤔 or why it’s more advantageous to use

lambda.invokeFunction

hey guys, how should I go about sharing a table construct across multiple stacks? Basically, I’ve an AppStack has my Table and then I’ve multiple Stacks which use Api constructs. I wan’t to addPermissions to the previously created Table in all these Apis.

I noticed

.map

files are being included when uploading code to Lambda. is there a way to prevent this, since I’ll (eventually) be uploading those to Sentry instead?

I have a core stack thats in its own "repo" (using monorepo rush.js ftw!). Core consists of VPC / Cognito User pool which allows me to control env (dev / sandbox / test / prod / whatever). I want to setup our BFF to us that userpool and create its own identity pool etc. (pretty sure this is a thing). Main goal is that every sandbox / dev / whatever doesn't have to kick up its own VPC / Cognito Pool / Maybe like redshift or something... unclear what else I'm going to put into here.

const userPoolId = cdk.Fn.importValue(`${props.infraCoreStage}-userpool-id`);
const userPool = cognito.UserPool.fromUserPoolId(this, CONSTANTS.environmentUserPool, userPoolId);

I can't seem to use the sst.Auth as userPool is a IUserPool.

this.auth = new sst.Auth(this, scope.logicalPrefixedName(CONSTANTS.authPrefixName), {
            cognito: {
                userPool, // does not like this
                userPoolClient // or this
            },
            identityPool: {
                allowUnauthenticatedIdentities: false
            },
            google: {
                clientId
            }
        });

1. Is this not a common pattern? 2. What horrible thing I'm I doing wrong? 3. I miss water cooler days too...

I am trying to create rabbit my using following construct.

const rabbitMQ = new CfnBroker(this, "name', {...});

I want to get connection string out of created rabbitMQ. So I used following code.

environment {
  RABBIT: rabbitMQ.attrAmqpEndpoints[0]
}

Even thought

attrAmpqEndpoints

is documented as

string[]

sst run stat

was throwing error saying I can access array using index directly and I can access it using Fn.select(index, list). So I changed code to

import { Fn } from '@aws-cdk/core';
...
environment {
  RABBIT: Fn.select(0, rabbitMQ.attrAmqpEndpoints)
}

Above code worked, but, I don't understand why there is need to use

Fn

for accessing array element at index. While at typescript should be able to use

list[0]

right?

Did somebody use this library to do authenticated HTTP calls to APIs hosted on AWS? https://github.com/mhart/aws4

anyone know how to get the identity pool ID associated w/ a Cognito user pool? it’s needed for Amplify Auth

any reason the esbuild

onEnd

callback doesn’t work/exist in SST’s environment?

{ onResolve: [Function: onResolve], onLoad: [Function: onLoad] }

Hey folks, I'm using a postConfirmation lamdbajs function, and it seems like when I deploy, imports/require statements don't work and cause an error. Code looks like this:

/* eslint-disable no-console */
const fetch = require('node-fetch');

export async function main(event, context, callback) {
  try {
    // Post user data to elastic

    console.log('posting data to elastic...', process.env.API_URL);

    await fetch(
      `${process.env.API_URL}/users`, {
        method: 'POST',
        headers: {
          'content-type': 'application/json',
        },
        body: JSON.stringify(event),
      },
    )
      .then((res) => res.json()) // expecting a json response
      .then(() => callback(null, event));
  } catch (error) {
    console.log('there was an error', error);
    callback(null, event);
  }
}

It works fine when testing locally, but after running

sst deploy

and trying out the workflow, I noticed this in CloudWatch:

2021-08-16T02:58:40.094Z	51cc72c5-0e44-4b15-8360-66170252a532	INFO	there was an error TypeError: fetch is not a function
    at Runtime.main [as handler] (/var/task/postConfirmation.js:1080:11)
    at Runtime.handleOnce (/var/runtime/Runtime.js:66:25)

So the one module I import is not working. It's definitely in my package.json:

"dependencies": {
    "@aws-cdk/core": "1.111.0",
    "@serverless-stack/cli": "0.38.0",
    "@serverless-stack/resources": "0.38.0",
    "node-fetch": "^2.6.1"
  },

Any thoughts?

Hey everyone, quick question. Does SST support env variables for auth or only the credentials file?

just wanted to check something: this is the correct way to handle dependencies that are binaries, correct?

bundle: {
      // sharp is a binary, and we have it in a Lambda layer, so exclude it
      externalModules: ['sharp'],
      loader: {
        '.node': 'binary',
      },
    },

whoa! SST/CDK just tried to do something really messed up. I have a route that has a JWT authorizer. I modified it to be:

'GET /media': {
    function: 'get-media.main',
    authorizationType: sst.ApiAuthorizationType.NONE,
  },

so I can test the endpoint without needing authorization. SST said “Press ENTER to redeploy infrastructure” and then this happened:

Deploying stacks
Checking deploy status...
dev-microservices-core | UPDATE_IN_PROGRESS | AWS::CloudFormation::Stack | dev-microservices-core
dev-microservices-core | UPDATE_ROLLBACK_IN_PROGRESS | AWS::CloudFormation::Stack | dev-microservices-core Export dev-microservices-core:ExportsOutputRefUserPoolClient2F5918F753847A55 cannot be deleted as it is in use by dev-microservices-api
Checking deploy status...
Checking deploy status...
dev-microservices-core | UPDATE_ROLLBACK_COMPLETE_CLEANUP_IN_PROGRESS | AWS::CloudFormation::Stack | dev-microservices-core
dev-microservices-core | UPDATE_ROLLBACK_COMPLETE | AWS::CloudFormation::Stack | dev-microservices-core

it tried to delete my Cognito user pool client??