Yes you can, or in the YouTube chat. We'll monitor both.

@white-balloon-205 I'm starting to use pulumi with AWS CodeBuild, which is basically a glorified bash prompt running in a container. How can i bring all that work into my pulumi script? E.g, building docker images, getting configuration, running arbitrary shell commands, etc.? How would I integrate that with Pulumi's "preview" vs. "live" run?

Any questions for @white-balloon-205 ? Can't guarantee a tshirt for questions but... well, actually that's about the only thing I can guarantee - so fire away.

how about a T-shirt if your name is mentioned? (he asks selfishly)

guaranteed to be worn at KubeCon 🙂

just curious, are you guys planning in the long term to create own providers(like the kubernetes one), instead of depending on the terraform ones?

thanks @white-balloon-205!

Thanks everyone for tuning in - let us know if you like this format, and we'll keep it going!

Howdy, I am trying to use the example https://github.com/pulumi/examples/tree/master/aws-ts-serverless-raw to point at a python Flask app instead of a C# app. I think this is possible? but I can't seem to configure the handler right and there are probably other issues... I am just using the simplest Flask app possible right now. Also Is this even how I want to do this?

Is there an example of pulumi managing the infrastructure while keeping the application code in separate repositories?

@fierce-ram-50276 that's how we are using it now

how do you connect the two? the only thing i can think of is packaging the app code into a docker repository and have pulumi connect to that

oh I see... yeah we aren't using Pulumi to deploy the apps for now at least

ohh that makes sense. I would like to be able to stand up a whole pulumi stack with a pulumi up, but i would like to keep the infrastructure a little separated from the app code itself.

ok

might be wishing for the impossible though haha

@fierce-ram-50276 Yeah - packaging up into a Docker Repository, or doing a Packer build of an AMI, or however else you build your deployable artifacts can happen in a deployment pipeline entirely separate from Pulumi - and then the Pulumi program can refer to the repositories or AMIs that are being published by that pipeline. Alternatively you can have two separate stacks, both in Pulumi, one which builds and pushes an image, and the other which references that image in Service. And then deployments of each can be triggered from the separate versioning contexts (perhaps pushes to two different repos).

yeah we have a separate pipeline for deploying the Helm charts. But we are in very early phases

thanks @white-balloon-205 and @orange-tailor-85423! I’ll keep digging

cool, please share what you come up with

@fierce-ram-50276 I wrote a few weeks ago how we’re doing things with pulumi roughly (we have multiple repos as well): https://pulumi-community.slack.com/archives/C84L4E3N1/p1538872835000100 . As an addition to what I wrote back then we also meanwhile have a small CLI tool (internal) which can checkout multiple repositories and pulumi up or pulumi destroy them all at once. That CLI tool is useful for the case when we want to spin up or tear down an entire environment.

thanks for sharing @glamorous-printer-66548! that sounds more along the lines of what I was invisoning in my head.

Peeps using Google auth......

TypeScript/JS is killing me

@orange-tailor-85423 What’s the problem?

LIterally every single example has some version of this:

but trying to code from the samples, that 2nd null is just not accepted

sorry, first null