Howdy all, I am trying to wrap my head around the responsibilities of

onCreate

/

onUpdate

/

onDelete

in the context of the

sst.Script

. I am looking to run DB migrations on each deploy. Do I have to mount the lambda against the

onCreate

and

onUpdate

properties? It is a bit confusing as the docs state that "It gets run on every deployment."

Hey guys, I get this error while testing WebSockets locally, works fine when running on lambda. I'm on v0.65.4 Thanks

Unhandled Promise Rejection     {"errorType":"Runtime.UnhandledPromiseRejection","errorMessage":"ReferenceError: require is not defined","reason":"ReferenceError: require is not defined","promise":{},"stack":["Runtime.UnhandledPromiseRejection: ReferenceError: require is not defined","    at process.<anonymous> (file:///home/sam/Desktop/wrs/node_modules/@serverless-stack/aws-lambda-ric/lib/index.js:34:23)","    at process.emit (node:events:394:28)","    at emit (node:internal/process/promises:136:22)","    at processPromiseRejections (node:internal/process/promises:242:25)","    at processTicksAndRejections (node:internal/process/task_queues:97:32)"]}

So more an inquiry, what SST constructor do we have that handles PUB/SUB requests? (outside of GraphQL if possible)

in SST, i know that we can use

auth.attachPermissionsForAuthUsers([api]);

when creating the sst.Auth() constructor. but what if i want to reference an existing already provisioned cognito userpool?

Is there a way to add a

*filterPatterns*

on a kineses stream? I think

addConsumer

does not support that but looking at an example in

SLS

it looks like it should be possible.

SOLVED:

Generally you should specify paths relative to SST.json

--- This might seem obvious but it isn’t to me so if someone knows, I’d appreciate it. I have the following directory structure:

src/
  lambda-handler/
    rewiter.js
stacks/
  SiteStatck.js
  index.js
sst.json

Inside

StaticSite

, I’d like to attach an edge lambda function. That looks vaguely like this:

const originRewriteHandler = new experimental.EdgeFunction(
      this,
      "OriginRewriteFunc",
      {
        runtime: lambda.Runtime.NODEJS_14_X,
        handler: "src/lambda-handler/rewriter.main",
        // this path is wrong and I can't figure it out
        code: lambda.Code.fromAsset(path.join(__dirname, "lambda-handler")), <----------------------
      }
    );

Everywhere I go with different paths I end up in

.build

Hello, When I run

sst deploy [stack]

it builds the whole functions and web components for all stacks. How can I just build the constructs related to the specific stack only?

Hey guys, anyone had any experience using the

Script

construct with VPC resources. Having ongoing woes with a post-deploy migrations script and have hit another wall:

getaddrinfo ENOTFOUND {rds address}

Any ideas? Thought it might be a permissions thing so I added a

PolicyStatement

to go allow all actions on the arn for the VPC and RDS instance. I’m stuck.

Hey there does anyone know how to run the

down

part of a migration using the new sst RDS construct? Can run

up

easily through sst console, but doesn't seem to be an easy way to clear changes if for example I need to change something while in development

We have an SSM parameter that is called at build. We've updated it. It's changed - really. It is. Yet, every time we attempt a deploy or build the old SSM parameter is used. We can't figure out where it's coming from. Has anyone else had this issue? Does SST cache SSM parameters?

Hi. I was stuck with this error CREATE_FAILED | AWS:Lambda:LayerVersion | PrismaLayer958161C6 | Could not unzip uploaded file. Please check your file, then try to upload again. (Service: AWSLambdaInternal; Status Code: 400; Error Code: InvalidParameterValueException; Request ID: 4fa5aa5e-24be-4cfa-83d5-ef466628c847; Proxy: null) and could not understand what was happening. But it turned out that the reason is that the S3 bucket where SST uploads lambdas and layers zip files is full. So I manually deleted all objects from the bucket and now the deployment works. Is there a way to remove files from SST service bucket (cdk-hnb659fds-assets-660857977215-us-east-1) automatically?

I've been trying to capture the output of

sst diff

into a variable within bash, but it appears the actual CDK diff is output via an inner child process or something. I end up getting something like this:

❯ body=$(./node_modules/.bin/sst diff --stage dev)
Stack dev-foo
IAM Statement Changes
... actual CDK DIFF

And outputting body variable:

❯ echo "$body"
Using stage: dev
Preparing your SST app
Building function src/foo/bar.handler
...

Anyone tried anything similar? I almost have a GitHub action working here. Just need to be able to nab the correct output. 😄

Hello all, I am using github action to deploy, actually it's taking about 40 minutes, and I want to add some integrations tests which requires running

sst start

so it will take about 1.5 hour to deploy. Actually when running deploy from local machine, it take much fewer time to deploy, I think because there is some caching process in sst I found that there is a caching option for github actions here: https://docs.github.com/en/actions/using-workflows/caching-dependencies-to-speed-up-workflows#using-the-cache-action So I want to ask what files and folders should I cache, is

.sst

folder enough?

Hopefully the last stupid question of the day: If I wanted all source to be included when a lambda builds (no tree shaking) how do I set this up? I assume i need to do something with the bundle.esbuild config right? @thdxr I imagine you may have some up against something similar sorting out migrations with Kysely?

I didn’t find anything in searching, so apologies if this has already been addressed. Is there any way to invoke EdgeFunctions locally using SST Console, or does that only work specifically with the SST Function construct?

the console.serverless-stack.com endpoint when starting a serverless-stack keeps stuck on "syncing metadata"

Having a strange issue where i have the AWS_PROFILE env variable set I can log it from within the stack however it this this error when it tries to deploy the nextjs site?

Need to perform AWS calls for account my-account, but no credentials have been configured
Found errors

I see for each ApiGateway route a role is created, is that right? If so how to reuse same role for each route.

i dont want to use a fully custom domain, just prefer to name something like "my-api-thingy.execute-api.us-east-1.amazonaws.com"

is there a DynamoDB client library for writing simpler queries? i could abstract some of it away myself but wanst sure if there was a NPM package that already does this.

What should I do to get the new AAAA records after upgrading to 0.65?

Just curious, hopefully I don't start a war here 🙂 Do folks test APIs by deploying to a test stage and running tests that hit those URLs on AWS through APIGateway or do you all like to mock AWS APIGateway events and call your functions directly?

Hi all, has anyone seen this error when building? SST version 0.63.0

EnvironmentCredentialsProviderFailure: Variable AMAZON_ACCESS_KEY_ID not set.

I'm only seeing it in the ci pipeline, where I've only set

AWS_ACCESS_KEY

and

AWS_SECRET_KEY

External dependencies — this is probably asked all the time. Is there an example of how to build a lambda layer that contains my node module built for my architecture? My steps so far: I added SSH2 as a dependency. I couldn’t deploy it. We flagged it as external. Everything deployed. Now I’m realizing that the lambda doesn’t actually have the module. 😞

for the seed CI-CD, how do i ignore "The engine "node" is incompatible with this module"

Hi, I'm trying to build an api service. For example a image background removal service (just for example) I mean, my users will have an api key which they have to include with every request. It will be used to limit the users quota according to billing plans. I'm planning to use supabase for database and auth. Is there any way I can do this with STT?

Hi everyone, I'm trying to build a stack that uses sst.RDS + sst.GraphQLApi and also Prisma as a database client. Prisma needs the database credentials like a URL connection but I don't know how I can get it. Maybe someone can help me please,

Hi folks, I've followed this guide to start using Aurora in an app I'm building - https://serverless-stack.com/examples/how-to-use-postgresql-in-your-serverless-app.html. This uses the

data-api-client

npm package to handle the requests and connection to the DB. I'm getting some timeouts when using this -

Failed to send response because the Lambda function is disconnected

I'm wondering can I use the AWS SDK RDS Data Client package instead - https://docs.aws.amazon.com/AWSJavaScriptSDK/v3/latest/clients/client-rds-data/index.html#aws-sdkclient-rds-data?

Made more sense to post this here: What’s the recommended project structure when using the 

sst.StackSite

 construct? I’m trying to deploy my CDK app with Seed, but I get an error when building the site because the node_modules for building the site are nested in another folder, and the build pipeline is trying to find it the project root

├── sst.json
├── node_modules // <- my SST dependencies
├── package.json
├── stacks
│   ├── index.tx // <- creates stack for deploying static site in /apps/static-site
├── apps
│   ├── static-site
│   |   ├── package.json
│   |   ├── node_modules // <- my React dependencies

The error I’m getting is 

react-scripts:not found

 (it’s installed in the 

static-site

 folder, not the project root) (edited)

Anyone ever seen this? “Parameter count 201 is greater than max allowed 200”