anyone know of a way to generate the esbuild metafile along with the build file? I see that it's option in NodejsFunction construct's

bundling

props, but I'm not seeing it in the SST function

bundle

props.

If I have a NextjsSite and run a

yarn sst deploy

then make no changes to anything and run

yarn sst deploy

again, it thinks the web app has changed and does an update. I assume that this is because a

deploy

executes a build and there must be something that is generated which changes every build even with no code changes? How can I prevent this? Is it a bug?

hey @thdxr could you bump this to 5.13.0 whenever you next do a push? https://github.com/serverless-stack/serverless-stack/blob/master/packages/eslint-config-serverless-stack/package.json#L11 4 still has a dependency on eslint 5 || 6 || 7, but SST is using eslint 8 and I am getting peer dependency errors. like this

Just curious if SST plans to support the Terraform CDK in the future or to be for AWS specifically?

Hey! Just getting back from updating sst 0.59 -> 0.67 (ultimate console experience sold it for me after a period of stability 😅) and I started looking for examples of migrating cdk.ServerlessCluster to sst.RDS construct. Obviously for a prod project I'd rather not lose my client's data even for a second. Anyone went through this topic and can recommend bulletproof steps? 😀

I’ve picked up and old JavaScript project that was deployed using ClaudiaJS and I’m converting it to use SST. In time, I need to rewrite it in TypeScript but for now I just need to get it up and running. I’m running into a problem when the endpoint is called which just says

ERROR build_failure: The function src/handlers/lambdaHandlers.bookAppointmentHandler failed to build

I can’t find any more information anywhere on what it was that caused the build to fail, where should I be looking?

Hey guys. Is possible to setup TTL on Dynamo for specific rows? For example, I have some records which are created on an

initial

state, which shouldn’t live forever… only on specific error scenarios that could happen. But on other cases, I don’t want to have TTL on some records, which has the

correct

state. Long story short…. I want to setup TTL for “some” rows and “some not” on the same Dynamo table.

hey @Frank (I think this is probably your domain, correct me if wrong), I am considering a construct to handle Next.JS Static exports, but handling the routing logic in Cloudfront for dynamic routes (

[id].tsx

). • Have you thought about this? • Do you think there is value and would SST have interest in such a thing? On the face of it, it would allow a simple, cheap entrypoint to Next.JS with capabilities more or less on par with something like CRA and Next.JS defaults and static exporting (not a SPA, but real static site). It would be devoid of features listed here.

Hello folks! For a new feature we were working on today we needed CDK’s aws_s3_deployment

Source.data

(see here). So we upgraded to the latest aws-cdk-lib. But it has some newer type definitions. For instance, the latest

s3.Bucket

type has a

withNotifications

function which is missing on the old type. This means that if you try to use `sst.Bucket`’s internal

s3Bucket

with other cdk APIs, it won’t type check. Do you have any plans to upgrade sst’s cdk-lib to latest? I wouldn’t mind helping. Thanks!

Is there a way to coordinate

npx sst start

for a local debug stack and

npm run dev

of a next.js site? I’m thinking for example when using turbo or other monorepo tooling that you can run a single command that starts the debug stack as a dependency to the next.js dev runtime. I can start them in parallel but the next.js dev process will fail because the debug stack has not yet created the files needed by

sst-env

. Alternatively, could

sst-env

rather than die and kill the process. The

sst-env

process could output a message to STDOUT, wait and retry until the file content exists. This way starting the processes in parallel would work and no additional process coordination would be needed.

Hey guys. We are still using Epsagon (planning to migrate in the near future). But something we noticed, is we stopped getting “database tracing” from RDS on the “timeline” view of a Lambda. Is somebody else using this feature? It was great to know how long the queries were taking and such, to monitor performance.

I noticed the S3 bucket that I assume SST uses to deploy, called something like

cdk-xx-assets-1234567-us-west-2

is not deleting old objects ... is there a reason why this bucket shouldn't have a lifecycle rule to remove old objects? Is there a recommendation on how to manage this bucket?

A couple of folks have noticed the S3 bucket in their CDK bootstrap stack is becoming bloated. I came across this discussion on why the old files are not automatically removed. And saw someone created a construct that creates a step-function behind the scene to periodically clean up unused files in the bucket.

With the migrations for RDS databases, if I create multiple database, how do the migrations know which database to run them against?

I have a few stacks which currently are deploying to

us-east-1

because of the

region

property inside

sst.json

. If I wanted to move all of them to

us-east-2

is there anything I need to do except change that property and deploy?

[Possible Live Dev Bug? `0.60.2`*]:* Saving one function deploys a completely different one.

When writing CDK, I find myself wanting to use environment variables locally (e.g. .env.local), but use ssm for long lived environments like stage/prod. Prioritizing env variables over parameter store values seems like a common pattern. Before I go writing a method to do this, is there a common recipe or util available for this purpose?

Interesting use case for SST Script? https://twitter.com/JannikWempe/status/1501835035070152705?t=XoEufaVRddymIfMcw4fqCg&s=19

Or better yet, the Lambda function infrastructure ie stuff declared outside the handler, is invoked once per instantiation, while the handler itself can then be invoked multiple times until the container is killed off.

How does Seed work if you only can deploy to subdomain.yourdomain.com? I have setup Route 53 to host our subdomain, but not the main domain. But when we build it says:

It seems you are configuring custom domains for you URL. And SST is not able to find the hosted zone "<http://yourdomain.com|yourdomain.com>" in your AWS Route 53 account. Please double check and make sure the zone exists, or pass in a different zone.

Do I have my stacks misconfigured?

I'm using a

NextjsSite

and I just upgraded from sst

0.60.4

to

0.69.0

and now I'm getting an error deploying (no changes to the stack):

| UPDATE_FAILED | Custom::SSTEdgeLambda | siteApiEdgeLambdaF3190F0F | Received response status [FAILED] from custom resource. Message returned: Missing required key 'FunctionName' in params

and I'm not quite sure where that's coming from Also is there a way to disable the Lambda@Edge functions if we don't use SSR?

Are there any plans to add docker/ecs stack? I'd love the option to deploy a gatsby/nextjs site using a docker container.

Hi all, I’m wondering about a best practice for managing keys and IDs generated by AWS and used in my mobile app, and how the app can use these values. Here’s the scenario: my mobile app uses Cognito for authentication to AppSync. It needs the user pool ID, a client ID, and an Identity pool ID. For unauthenticated requests to AppSync, I use an API key. However, if I ever have to regenerate my production stacks, all of those keys will change and the mobile app will break because they are currently stored in a file in the mobile app. How can I prevent this brittleness? I think it’d be nice to have an API endpoint that the client app calls upon opening to download those values. Should those values be stored in a Secrets Manager, read by a lambda, and sent to the client? If you’ve got any ideas, I’d really appreciate it! Thanks!

this is becoming very urgent as the web platform I'm working on is currently facing issues

Has anyone managed to get break points working with SST-Golang?

Also to @Frank and all the SST Folk… I’ve bumped into other teams at my company, Analog Devices, using SST just over the last few months. Just randomly they discovered it independently. It seems to really be catching on !!!

For people using typeorm, how are you setting it up and how does the connection work? I’m just getting

TypeError: Cannot read property 'filename' of undefined

whenever I try to make a connection

This morning I tried using the console for the first time. And it's stuck on "Syncing metadata" At first (reading stuff here) I thought it was the Brave shields but I tried in chrome and same result... 403 errors as per the browser network screenshot :

Just want to say THANKS! again to the SST team, this time specifically for your SocialCards example. We took this and created a simple reusable construct that we can now implement for any our our sites (we have many) that need this functionality. Thanks!

Are you guys going to add support for Gatsby at any point? I'm currently using the static site stack for it at the moment.