set the channel description: Channel for general questions about Pulumi

set the channel topic: Ask any question about Pulumi

set the channel topic: Ask any question about Pulumi!

/poll "Tabs or Spaces?" "Tabs" "Spaces"

@chilly-crayon-57653 pinned a message to this channel.

@polite-receptionist-39388 Welcome! You’re our first beta user on pulumi-community, which is why it’s a bit empty. The entire team is here, so feel free to ask any questions or provide feedback on your experience so far.

no problem at all! happy to be here

big fan of @big-piano-35669’s work 🙂

Hey y'all! Super excited about the possibilities with Pulumi and this concept of mixing and matching infrastructure code with application code. Just got done reading over some more examples and the docs. Couple of thoughts: 1. local testing would be really nice - perhaps hooking up to aws localstack for deploying locally, or providing some testing framework to be able to validate pulumi code off-line so that the dev cycle is sped up. 2. reliance on central pulumi service for storing stack state (and therefore being blocked in deployment if it goes down) would be a show-stopper for any of the clients I work for - saw some discussion in the other slack about this and glad to hear it's being discussed. My thought is I would love to see a config option for where I store my stack state - i.e. being able to specify an S3 bucket/dynamodb for locking the way terraform does it

Thanks again for including me in your private beta!

Hey Will - welcome, and thanks for the great feedback! On local testing, we're definitely interested in doing more here. One of the nice benefits of building abstractions using Pulumi over top of your cloud resources and then deploying applications authored in terms of those is that the abstractions can be easier to mock and test. We actually had the beginning of a mock layer for

@pulumi/cloud

, which did exactly that - but it had fallen a bit behind so we recently removed it - we're fully expecting to bring something back here in the future - either based on something like

localstack

, or running against a local Kubernetes cluster. But more importantly, you can always build your own abstractions that are mockable for your own workloads. Definitely interested in exploring this more if you have more thoughts after spending some time using Pulumi. Definitely hear the feedback on the central service for storing stack state. We're thinking a lot about how best to offer additional options here. Would love to do a deeper dive with you on what requirements you expect the clients you work with would have, and how we could meet those requirements while still providing a robust-by-default experience for users.

@worried-agency-74630 On the question of running locally: for your scenarios, is it important to emulate the entire stack, or just the compute? For instance, I tried a tool recently that ran a Lambda function on my local machine, but it still connected to my actual Dynamo instance in AWS.

Our latest live stream is available to watch. @white-balloon-205 walks through building an app with Pulumi from scratch using lambdas, containers and

bucket.bucket.bucket

Heads up on a fun Twitch session with Wednesday with @big-piano-35669 - feel free to ask questions in advance or at the time. https://twitter.com/PulumiCorp/status/1006284943251689473

Hey, I'm playing around with the TypeScript SDK for Azure

I'd like to setup an App Service that uses Azure SQL, storage accounts/containers, and Azure Search

I can configure my App Service to use the storage Account by passing the

primaryAccessKey

to the App Service

However, Azure Search and Azure SQL don't seem to expose their credentials

@flaky-continent-74538 To clarify, are you using the

@pulumi/azure

package?

Yup

How can I get the Azure Search's API key?

And how can I get the Azure SQL database's connection string?

I see, those properties are missing from the API. I suspect you may have to use runtime APIs by using the Azure JavaScript SDK, but let me check with my teammates to see if there’s a better solution.

Ah okay. Thanks

I think I figured it out for Azure SQL. You have to create a

sql.SqlServer

object to create your

sql.Database

object. The

SqlServer

has properties for the domain name and login credentials. This isn't every evident from the

Database

type alone.

Still trying to figure out Azure Search 🙂

Our SDK is built on top of the Terraform Azure provider. So, you might be able to find some documentation that describes how to do it in Terraform.

loic ... can you use Azure.core.TemplateDeployment ... this lets you use snippets of ARM

//something like:const template = new azure.core.TemplateDeployment(deploymentId, { deploymentMode: "incremental", resourceGroupName: resourceGroup.name, templateBody: JSON.stringify(currentBillingFeatures.jsonTemplate), parameters: { "deploymentId": deploymentId, "appName": insights.name, "appLocation": resourceGroup.location, }, });

except insread of billingfeatures, you use the ARM stuff for search