I'm curious if anyone has already setup a Kinesis Data Firehose using SST ? I'm simply looking for a straightforward way to log audit data into S3 and if needed query it using Athena later down the road. Didn't find any example for this.

Hey all! I am curious the the current state of Serverless NextJS. Does it still have the stack teardown issue of using Lambda@Edge and the fact that they can take up to 2 hours to delete? Is anyone familiar with this? Currently this messy cleanup is the biggest issue prevent us from using sst

NextjsSite

I'm using the CDK Kinesis Firehose construct as below but it doesn't manage to deploy the stack giving me a "Dynamic Partitioning Namespaces can only be part of a prefix expression when Dynamic Partitioning is enabled." error when trying to create the AWS:KinesisFirehose:DeliveryStream . I tried googling this without success and can't find where in CDK I could enable dynamic partitioning. Any idea anyone ??

import * as firehose from "@aws-cdk/aws-kinesisfirehose-alpha";
import * as firehose_destinations from "@aws-cdk/aws-kinesisfirehose-destinations-alpha";
...
    const auditBucket = new Bucket(this, 'Audit')
    const cdkAuditBucket = auditBucket.cdk.bucket
    const firehoseDestination = new firehose_destinations.S3Bucket(cdkAuditBucket, { dataOutputPrefix: 'site!{partitionKeyFromQuery:id_site}/!{timestamp:yyyy}/!{timestamp:MM}/', errorOutputPrefix: 'firehoseErrors/' })

    const kinesisFirehose = new firehose.DeliveryStream(this, 'Audit Firehose', { destinations: [firehoseDestination] });
    config['AuditKinesisDeliveryStreamName'] = kinesisFirehose.deliveryStreamName

Is there a way to put the

.env

files into a directory, instead of keeping them in the root? The underlying library supports a

path

option: https://github.com/motdotla/dotenv#path As there are a bunch of different

.env

files, it seems cluttered to me to store them all in the root, instead of a

dotenv

or

config

directory of some kind. Ideally I'd be able to add

dotenv.path

to

sst.json

, I think?

Hi 👋 I'm running into some issues with migrations, I'm receiving the following error:

e8a5a024-b37d-4ce6-b85e-d0d7ee454267 REQUEST dev-dolartec-app-db-DatabaseMigrationFunction9CBE9-qOutRhPjrpC9 [index.handler] invoked
(node:19518) Warning: To load an ES module, set "type": "module" in the package.json or use the .mjs extension.
(Use `node --trace-warnings ...` to show where the warning was created)
e8a5a024-b37d-4ce6-b85e-d0d7ee454267 ERROR Runtime.UnhandledPromiseRejection: Runtime.UserCodeSyntaxError: SyntaxError: Cannot use import statement outside a module

I’m wanting to create a single RDS cluster which then has multiple databases inside (one for each of my microservices) but I’m stuck at the

migrations

being a string to a path. I need to be able to know which database is targeted for migration and then set the path for the migrations accordingly. Something like this

export async function RDSStack({ stack, app }: StackContext) {
  const cluster = new RDS(stack, "RDSCluster", {
    engine: "postgresql10.14",
    defaultDatabaseName: 'postgres',
    migrations: (dbname: string) => new FileMigrationProvider(`./migrations/${dbname}`)
  })

  return { cluster }

@Frank @thdxr is this something that could be looked at or is there way that I can tell the underlying Kysely to use a custom

MigrationProvider

class?

Hey guys, I have a bucket declared like:

return new Bucket(this, 'my-bucket', {
      bucketName,
      cors: [
        {
          allowedOrigins: [allowedOrigin],
          allowedMethods: [HttpMethods.GET, HttpMethods.HEAD],
        },
      ],
      lifecycleRules: [
        {
          expiration: Duration.days(1),
        },
      ],
    });

Now I want to remove the TTL… I removed from code, but CFormation is trying to re-create the bucket and fails… why can’t just UPDATE it?

Howdy ya'll. I've updated the PR for the

RemixSite

construct to contain a description. This should hopefully help with reviews. https://github.com/serverless-stack/serverless-stack/pull/1800

Is there a way to disable the sst console?

@Frank I was checking the source for

StaticSite

, due I have to deploy multiple React apps into the same bucket… (for the micro-frontend discussion)… And I need to copy the logic for invalidate CFront… but… why is the invalidation done in a lambda, and not done from the stack? is that even possible?

This message was deleted.

Is anyone using graphql-codegen with a GraphQLApi behind an authorizer?

RemixSite

PR update; ✅ Added tests

I’m at sst@0.69.7 and I’m planning to upgrade to v1.2.25, anything that I need to concern myself with other than the upgrade guide itself?

I've added custom domains (Route53) to my api's and static site across diff stages - I've deployed changes across my stacks but its' been stuck on checking deploy status... for over 5 hours now. Is this normal? I guess it takes a while provisioning ssl certificate.

I am sure this has been asked before. I’d like to separate my stacks in multiple repos (for compliance / access control) but deploy to the same account / stage (for instance my sensitive credentials stack is separated in another repo but I need to deploy to stage prod with the same credentials. Is there any caveat / considerations ?

Hi there, I decided to choose SST as my primary framework for building big new monorepo project. I have 5+ years of validated experience with building big monorepo project based on Serverless Framework, but I wanted to upgrade to something more modern. I started to prepare the monorepo and I stumbled upon a many issues and questions, but now I want to discuss the one which would probably require a code modification of SST. Use case: I have a monorepo managed via NPM workspaces feature. I want to import a file from other package

import { someFunction } from '@packages/some-package' // or import from '../../packages/some-package' - it's is the same in this case

export const handler = async (event) => {
	someFunction('hello')
    return '123'
}

Then I start the debugging session with

$ sst start

When I change the file of the lambda function handler, I can see in the terminal info message:

Functions: Building api functions/lambda.handler...
Functions: Done building api functions/lambda.handler (101ms)

And everything works great so far. 👍 The problem: When I change a file inside the package

@packages/some-package

which is located in completely different folder (

../../packages/some-package

), no reload/recompilation happens and the application is still using the old source code of the package. How I hacked it and fixed it: When I did the reverse engineering, I found out that here: https://github.com/serverless-stack/serverless-stack/blob/9d8cc00da72942ef39348b25ff993117e502f30f/packages/core/src/runtime/watcher.ts and here: https://github.com/serverless-stack/serverless-stack/blob/9d8cc00da72942ef39348b25ff993117e502f30f/packages/core/src/runtime/handler/node.ts#L2[…]24 the SST is defining which files it will be watching. When I added to the list path to my packages folder, it started to work ✅ What do you think would be the best way how to address this issue? I’m happy to contribute. 🙂 Thanks.

I have similar issue @Emanuel Bohnenkamp did you manage to fix it?

👋, can I get some 👀 and feedback on this PR: https://github.com/serverless-stack/serverless-stack/pull/1822 ? Thanks!

Hey, I have a problem with Live Lambda Debug. I'm using the launch.json provided in docs: https://docs.serverless-stack.com/live-lambda-development#debugging-with-visual-studio-code But when I Debug SST Start in VSCode it gets stuck indefinitely on Deploying Stacks... and I'm unable to hit any endpoints / routes and app falls over. Any ideas?

Hi all! I want to reinstall SST app on PROD but I want to preserve current buckets. Is there a way to do this without of seeing error

"a bucket with this name already exists"

?

does SST have any special support for Step Functions?