Serverless Stack

is there any way to speed up this process? something we can do on our end in configuration?

So at the moment...no. This is a quirk with CDK where everything is JS constructor based which means everything is forced to happen synchronously

My plan is to spend some time studying turborepo and see if we can trick CDK to get around this constraint

got it. I didn’t know CDK bundled/minified JS code.. I thought it was purely IAC?

or is the “Building Lambda function..” something else

CDK is basically a compile to Cloudformation framework. So it has a "synthesize" step where it compiles your cdk code into cloudformation code. At that point it needs to know where the function assets are so we build it there

<@U01JVDKASAC> can you think of a way to give CDK a fake function and then replace it at the end? It would allow us to parallel build some of this stuff + take advantage of caching strategies

I guess we could manipulate the raw cloudformation

<@U01MV4U2EV9> isn’t the “fake function” kinda like what serverless framework do when they create stack with S3 bucket only first, then update stack with actual infra included in the stack… Maybe you could do something similar, where you create a “hello world function” then replace, but not sure if it would be faster at all

Totally doable! Yeah, we’d have to manipulate the raw CFN, but that’s pretty straightforward

<@U01MV4U2EV9> is esbuild taking long in this case? or is the spawning new process part?

<@U01JVDKASAC> its everything. Spawning one by one + forced to spawn a new node proc

Got it! Created an issue here <https://github.com/serverless-stack/serverless-stack/issues/1176>