How can I "_hot reload_" my lambda functions using Pulumi? I have been using

aws.lambda.Function

to deploy to my localstack setup and I was hoping I could have a fast way to iterate over local changes, but I couldn't find a way to do so. I noticed that

aws.lambda.Function

zips the code (

pulumi.asset.AssetArchive

), which means that enabling hot swapping in localstack won't make any difference unless we have a way to make Pulumi repack the content of the lambda on demand. • I have tried using pulumi watch, but with no success. It seems that only the pulumi code is being watched, not the code of the lambda functions. • I have also tried running

pulumi up

every time a change was made on the lambda code, but that has extreme bad performance and possibly other side effects. What should I be looking at? Are there workarounds I could be trying?

What are the performance issues you’re seeing when you run Pulumi up? What are you using as a back end for the state?

@brief-alligator-51254 I gave me two cents for a solution in the discussion thread: https://github.com/pulumi/pulumi/discussions/11065#discussioncomment-3937307

What are the performance issues you’re seeing when you run Pulumi up?

I would say that

pulumi up

is just not designed to perform such action. The

Previewing update

phase takes 39s and the actual

Updating

phase takes 47s. The performance is comparable to running

pulumi up

in a clean state, so no hot reload by definition.

What are you using as a back end for the state?

Local Filesystem

I gave me two cents for a solution in the discussion thread

Thanks @colossal-diamond-64391!

So couple of things would help us here: 1. Can you run

pulumi up --logtostderr --logflow -v=9 2> logs.txt

2. Can you run (you might want to destroy the stack before you run it):

pulumi up --tracing=file:./up.trace

And then send over the

logs.txt

and the

up.trace

file that you produced to me: piers@pulumi.com Just as an aside, Pulumi does write a lot of checkpoints and if you’re using the local file system as your backend, especially on your local computer, you might find it’s not as performant as using the Pulumi service.