pulumi looks awesome! For my project, I have been using the serverless framework targeting AWS with APIGateway, Lambda, and DynamoDB. I've also been using serverless-offline and serverless-dynamodb-local plugins for rapid development and debugging. Is there a plan to support offline mode in pulumi? I'm curious to know how to reference a dynamodb table in a function that can be cloud-based or local.

lets say I have a golang serverless function I want to want to deploy to AWS, golang functions need to be compiled before uploading, does pulumi provide this capability (similar to the docker build capability)

Howdy!

does pulumi have a ui ?

Is there any way to share stack / stack state with other developers without creating an Organization in pulumi?

i just use a separate git repo for state

pulumi --local file://path/to/git/repo or something along those lines

sets the state folder

Does anyone have any examples of using GCP InstanceTemplate with startupScript?

Nevermind, got it 🙂

What does this mean?

Big thanks to Wallaroo's @bright-motherboard-93404 for this post earlier this week: using a little Pulumi power to enable on-demand analytics clusters: https://blog.wallaroolabs.com/2018/10/spinning-up-a-wallaroo-cluster-is-easy/

Hey all - we'll be at Hashiconf next week. If you're there, and would like to meet, you can find us at Booth S14 in the Pavilion Room. If you want to book a specific time slot, you can hit this link: https://info.pulumi.com/meetings/team-pulumi/hashiconf-catchup

Due to lack of TypeScript and general programming... what if I create my GKE cluster. And now I want to create namespaces inside of it

how do I reference the cluster I created as an input for the creation of the namespace object

Does pulumi support the full k8s resources definitions ? not only deployment files but statefulsets etc .. ?

i cant figure what is supported/not supported from the docs

Can i tell pulumi to fetch the configs from a git repo url directly ? in a multi repo case for example ?

I’m getting the following error:

Cannot call '.get' during update or preview

and I think it’s from the following new TS I added to my index.ts:

let apiIntentRobotServiceAccount = new k8s.core.v1.ServiceAccount(
  'api-intent-robot-service-account', {
    metadata: {
      name: 'api-intent-robot',
      namespace: context.inferredNamespace
    }
  }
)

let apiIntentEphemeralRobotRole = new k8s.rbac.v1.Role(
  'api-intent-ephemeral-robot-role',
  {
    metadata: {
      name: 'api-intent-robot-role',
      namespace: EPHEMERAL_NAMESPACE_NAME
    },
    rules: [
      {
        apiGroups: [''],
        resources: ['pods', 'pods/status'],
        verbs: ['get', 'create', 'patch', 'delete', 'list', 'status']
      }
    ]
  }
)

let apiIntentEphemeralRolebinding = new k8s.rbac.v1.RoleBinding(
  'api-intent-ephemeral-robot-rolebinding',
  {
    metadata: {
      name: 'api-intent-robot-rolebinding',
      namespace: EPHEMERAL_NAMESPACE_NAME
    },
    roleRef: {
      name: apiIntentEphemeralRobotRole.metadata.get().name,
      apiGroup: apiIntentEphemeralRobotRole.apiVersion,
      kind: apiIntentEphemeralRobotRole.kind
    },
    subjects: [
      {
        kind: apiIntentRobotServiceAccount.kind,
        name: apiIntentRobotServiceAccount.metadata.get().name,
      }
    ]
  })

Am I not allowed to get the name of a resource to use in the spec of another resource?

How feature complete is the go client as opposed to the js code ?

Meaning am i better off using go or js ?

Hey Folks, any thoughts on https://github.com/pinax/django-user-accounts/pull/293 ??

If i want to provide a configuration file to override some of the fileds like for example the number of replicas in a k8s deployment file how do i pass that to

pulimi up

I got pulumi into a weird state where it thinks certain resources exist in my k8s cluster when they don’t. Is there a way to ask pulumi to forget about these resources?

pulumi refresh

?

👋 I'm testing the waters a bit and am trying to create a couple services in AWS. Are there any downsides to pulumi spinning up an ELB per service rather than using host-header routing on a single balancer?

ELBs are quite expensive

right.. that's what I was kind of seeing