pulumi-community #general

millions-furniture-75402

05/21/2020, 3:59 PM

I’m trying to setup default behavior for a load balancer listener, but I run into an issue with defaulActions. Can someone shed some light on what I might be doing wrong?

ecs-with-lb-default-rules.ts

best-lifeguard-91445

05/21/2020, 4:48 PM

How can I create a

aws.sqs.QueuePolicy

with a dynamic list of statements? I'm getting: Calling [toString] on an [Output<T>] is not supported.

Copy code

function createQueuePolicyStatement(inboundSqs: aws.sqs.Queue, topicData: TopicData[]): pulumi.Input<string> {
    const statement = (queueArn: Output<string>, topicArn: Output<string>) => {
        return queueArn.apply((qarn) => {
            return topicArn.apply((tarn) => {
                return `
                        {
                            "Effect": "Allow",
                            "Principal": "*",
                            "Action": "SQS:SendMessage",
                            "Resource": "${qarn}",
                            "Condition": {
                                "ArnEquals": {
                                "aws:SourceArn": "${tarn}"
                            }
                        }`
            });
        });
    }
    const policyStatement = `
        "Version": "2012-10-17",
        "Id": "sqspolicy",
        "Statement": [
            ${topicData.map((data) => statement(inboundSqs.arn, data.topic.arn)).join(`,`)}
        ]
    }`;
    return policyStatement;
}

dazzling-memory-8548

05/21/2020, 6:49 PM

I have a question about lambda closure serialization. We're using nodejs with pulumi to deploy our lambdas as callback functions (https://www.pulumi.com/docs/reference/pkg/nodejs/pulumi/aws/lambda/#CallbackFunction). We recently pulled in a dependency (

parquetjs

) that has a compiled sub-dependency (

lzo

). We've now noticed that every time we rebuild our

node_modules

directory (with

npm ci

), the hash of the compiled

lzo

module changes, causing pulumi to want to redeploy all of our lambdas (even though no deps have actually changed). Is there an elegant workaround to prevent pulumi from continuously redeploying our lambdas? We've already tried excluding

lzo

codePathOptions

. cc @worried-engineer-33884

wonderful-dog-9045

05/21/2020, 7:01 PM

Are there any URI arguments to https://get.pulumi.com/? Like version, install directory, etc?

cool-egg-852

05/21/2020, 8:05 PM

Does a pulumi ComponentResource not share it’s

provider

with it’s children? I’m having an issue with a setup such as: Monitors (ComponentResource) (provider set) Monitor (ComponentResource) datadog.Monitor The

datadog.Monitor

doesn’t seem to have it’s

provider

set as far as I can tell. I was under the impression that the

provider

should be on it.

best-lifeguard-91445

05/22/2020, 3:15 AM

Does anyone have an example of creating an Aws Aurora cluster with read replicas? I have a cluster, and cluster instance, and cluster end point defined, but I don't see how to tag one of the instances as 'writer' false (it's a readonly output).

kind-mechanic-53546

05/22/2020, 4:42 AM

Is it possible to unsecretify an output? Building a docker image, passing a secret to it, results in the image name then being a secret as well, not necessary and somewhat annoying

green-salesclerk-34937

05/22/2020, 6:19 AM

Does anyone have tried https://github.com/pulumi/examples/tree/master/kubernetes-ts-jenkins ? It looks simple but I’m not getting through it. with some errors like attached image, but I’ve got through prerequite eks-nginx quite smoothly.

green-salesclerk-34937

05/22/2020, 6:20 AM

I’ve set enough

maxSize

in cluster opt

gray-carpet-67751

05/22/2020, 8:12 AM

Hey guys 👋 I am getting issues trying to select my stack

Copy code

2020-05-22T08:04:50.9212982Z error: no stack named 'dfds-frontend/production' found
2020-05-22T08:04:50.9944971Z error: --yes must be passed in to proceed when running in non-interactive mode
2020-05-22T08:04:51.0637194Z error: --yes must be passed in to proceed when running in non-interactive mode
2020-05-22T08:04:51.5013922Z error: no stack selected; please use `pulumi stack select` to choose one

This stack actually exists and it worked like a week ago but for some reaso it's been failing for a few days. Any pointers you could give me?

prehistoric-gold-59191

05/22/2020, 9:54 AM

Hi, need a little help here. We are looking to build self-service into our app. Does the Pulumi community version has the support for the self-service API?

incalculable-army-40859

05/22/2020, 2:08 PM

Does anyone know what version of Helm Pulumi uses? I am using Typescript and there only seems to be a

helm.v2

module. Does use Helm 3 or Helm 2?

clever-plumber-29709

05/22/2020, 5:26 PM

Hey guys, question about Inter-Stack Dependencies, and somewhat related the stack outputs. where are this values actually stored? or how do pulumi access them? I ask because i'm unsure if they are stored at the state or the config files, or in some way in the environment after the execution. I'm thinking about cases with a Micro-Stacks design, if I haven't run the infra stack on my machine, can i run the services stack that references it? or do i need to run the infra first, even if it doesn't change anything? Or in case one member of the team runs one micro-stack in one end, and another member of the team need to run another micro-stack that depends on it, in another end

best-lifeguard-91445

05/22/2020, 6:03 PM

after after running

pulumi up

successfully, subsequent runs are producing this error:

Copy code

~   ├─ aws:rds:Cluster              rds-cluster-patient-access-prod  **updating failed**     [diff: ~availabilityZones]; 1 error
     ├─ aws:autoscaling:Group        app-services                     **failed**              1 error
     └─ aws:autoscaling:Group        edge-services                    **failed**              1 error

Diagnostics:
  aws:autoscaling:Group (app-services):
    error: transport is closing

  aws:autoscaling:Group (edge-services):
    error: transport is closing

  aws:rds:Cluster (rds-cluster-patient-access-prod):
    error: transport is closing

  pulumi:pulumi:Stack (aws-prod):
    error: update failed

    panic: fatal: An assertion has failed: Expected diff to not require deletion or replacement during Update of urn:pulumi:prod::aws::aws:rds/cluster:Cluster::rds-cluster-patient-access-prod

I'm not making any changes to the cluster or related resources (availability zones, auto scaling groups).. any ideas?

billowy-nightfall-59212

05/22/2020, 6:48 PM

Hi, I am looking for engineering opportunities. I have built with whole cloud/k8s infrastructure with

pulumi

for more than a year. Please DM, if you are hiring. Thanks

chilly-hairdresser-56259

05/22/2020, 9:05 PM

Is anyone else out there having a rough time getting Pulumi to use aws Profiles? I have a shared credentials file in ~/.aws/credentials that has a default profile which has my access key and secret then I created another profile named sandbox which references a role_arn and source_profile of default. In the pulumi config I set aws:profile to use sandbox. However when I do a pulumi preview and write out the callidentity is still pointed to my main account.

best-lifeguard-91445

05/22/2020, 9:41 PM

Is there a way to set the ECR repository name without the random suffix? deleteBeforeReplace isn't an available option, nor is 'name'.

great-byte-67992

05/23/2020, 4:36 AM

Has anyone seen the new gitlab 13 feature GitLab HTTP Terraform state backend. I’m interested if pulumi could implement a compatible state backend for it and any other terraform http state backend. https://about.gitlab.com/releases/2020/05/22/gitlab-13-0-released/#gitlab-http-terraform-state-backend

witty-spring-28765

05/23/2020, 9:54 AM

Hello All, I am trying to acheive the following. Please let me know if this is possible 1. Create a python docker using pulumi 2. Deploy this docker image on to an EC2 instance can this be done is the same script or is it better to use a different CI/CD for the docker build and a separate one for Infra build.

able-beard-29160

05/23/2020, 11:43 AM

Hi, folks! Can stacks be used to run different code paths (TypeScript in my case) or would I need different projects for that? Ideally I wouldn't have to query the stack name in my code and executed based on that, but if that's the only solution, I'd like to know.

wonderful-dog-9045

05/23/2020, 11:48 AM

is the nodejs sdk version always equal to the pulumi runtime version? i.e. npm @pulumi/pulumi@2.2.1 -> pulumi version v2.2.1 Or is it coincidental that the current sdk version number is the same as the current runtime version number?

able-beard-29160

05/23/2020, 6:04 PM

Does

pulumi.StackReference

work with anything else other than the Pulumi service? I'm not sure what to pass as a stack name to the constructor.

green-school-95910

05/24/2020, 4:22 AM

Hey Pulumi team! Where should I send suggestions regarding the console of the Pulumi service? It would be nice if the resource details page showed the URN of the resource. Just now we had a possible credential leak and had to regenerate that key using the

--replace

option and the only way I could find to get the urn of a resource without having to change the code was to inspect the state using

pulumi stack export

green-school-95910

05/24/2020, 4:32 AM

Also, secondary suggestion: On the timeline tag of the update activities resources that are replaced could appear 2 times, one for when a new one is created and another for when the previous resource is deleted. It is strange how it appears today. The key needs to be replaced for the provider to change, but since it is only completely replaced after its dependency are updated it shows up after it, which can be confusing if someone needs check the operations at some point (auditing for example, it is not clear when the new resource was created).

astonishing-gpu-37274

05/24/2020, 7:36 AM

I'm getting this error for any lambda I'm trying to create:

Copy code

error: Error creating Lambda function: ResourceConflictException: Function already exist: ****-8d56ff7
    {
      RespMetadata: {
        StatusCode: 409,
        RequestID: "a4e2508b-7d19-43cc-93db-e08b1f64914e"
      },
      Message_: "Function already exist: ***-8d56ff7",
      Type: "User"
    }

I tried using --refresh, recreating the stack, switching to a different account. The lambda is actually created even when this error occur Looking at the AWS cloudtrail events, I see that CreateFunction is called twice, with a 2 min difference(the time it takes to create it). Using the latest pulumi version, anyone knows how to resolve? As a workaround I'm quickly deleting the lambda before the 2nd time it creates it, and that seems to make the pulumi update finish successfully

blue-businessperson-90950

05/24/2020, 8:52 AM

How is pulumi-azure related to the azurerm terraform provider?

blue-businessperson-90950

05/24/2020, 8:53 AM

It seem that pulumi use the terraform provider under the hood. It's true?

blue-businessperson-90950

05/24/2020, 8:54 AM

If it's true, is there a plan to remove the dependency?

alert-restaurant-79151

05/24/2020, 2:30 PM

What would be required to implement a provider in a new language (not golang)?

brainy-psychiatrist-41826

05/24/2020, 5:37 PM

using github actions, is there a way to run pulumi for a go project and not re-download the deps each time? i vendored the deps via

go mod vendor

, but the GH action always refetches the deps