- task: Pulumi@1
      inputs:
        azureSubscription: 'Test (bla-bla)'
        command: "up"
        cwd: "MyProject.Infrastructure/"
        stack: 'dev'
        args: '--yes'
    - script:
        echo "##vso[task.setvariable variable=test;isOutput=true]$(pulumi stack output test)"
      displayName: 'Set stack outputs as variables'
      name: 'pulumi'
    - bash: |
        echo "Using an input-macro works: $(test)"

Hi peeps, a question. If I run

pulumi stack output test

it gives me the output I'm expecting, and I can see in the logs that the pulumi outputs are correctly being registered

but when I try to output them it seems like they're not being set properly

any thoughts?

ok, the script that outputs the variables needs to be in the correct workingDirectory (which in this case is MyProject.Infrastructure)

good evening, I am trying to create a Kubs

PersistentVolumeClaim

object and then refer its name in another object (a

CronJob

)

sadly I can't refer to the

resource_name

I used because Pulumi changes names... which is fine. But how to get the name? I the

.id

field of the resource almost get me there but it includes the namespace in there. I just want the "new" name.

One last question: we did our Docker image build using a deploy.sh script before, and we could get some great build performance using the simple

--cache-from

argument when calling

docker build .

. We'd usually do something like

--cache-from=<http://registry.com/dir/image:latest|registry.com/dir/image:latest>

and then update the latest tag to the newly built and pushed image.

Right now I am using

class Image(pulumi.ComponentResource)

which has nothing that leverages caching

any idea on leveraging this --cache-from property while keep Pulumi in charge of our image building?

Kinda curious, I used pulumi to bootstrap an aws typescript project and when it didn’t find my credentials the error included a link to a terraform help guide, heh. Are some pieces of pulumi built on top of terraform?

Pulumi thinks it needs to replace my

aws:ecs:TaskDefinition

on every

up

run even though nothing has changed. Looking at the diff output for the containerDefinitions parameter on the TaskDefinition, I can see the only difference is on both my containers,

mountPoints

,

portMappings

,

user

, and

volumesFrom

currently exist as empty arrays (except user which is

"0"

) and do not exist in the desired TaskDefinition. I am not setting these properties.

Does the

--show-reads

property not work correctly for diffs? Whenever I do a

preview --diff

it shows reads without that flag set.

Hi all, is it possible to create a single stack that spans multiple cloud environments? I need to create a vpn connection between AWS and Azure. Would be nice if I could put the code for both ends of this setup in a single stack.

Can one shorten the random suffix on the resource names? Maybe with

ResourceTransformations

(dotnet azure sdk)? Thanks

I’m having a real challenge figuring out the right way to validate a certificate which has multiple alternative names - That will result in multiple DNS validation records being generated.. i think this will handle that, but i want the resulting

CertificateValidation

resource to be used in a

Depends

later in the program so that the Cloudfront CDN isn’t created before validation is complete…

s.cert.DomainValidationOptions.ApplyT(func(opts []acm.CertificateDomainValidationOption) (*acm.CertificateValidation, error) {
                var fqdns pulumi.StringArray
                for i, opt := range opts {
                        certValidationDomain, err := route53.NewRecord(ctx, fmt.Sprintf("%s-%d-validation", s.hostname, i), &route53.RecordArgs{
                                Name:    pulumi.String(*opt.ResourceRecordName),
                                Type:    pulumi.String(*opt.ResourceRecordType),
                                Records: pulumi.StringArray{pulumi.String(*opt.ResourceRecordValue)},
                                ZoneId:  s.zoneId, // zone.ID(),
                                Ttl:     <http://pulumi.Int|pulumi.Int>(600),
                        })
                        if err != nil {
                                return nil, fmt.Errorf("create route53 validation record failed: %v", err)
                        }
                        fqdns = append(fqdns, certValidationDomain.Fqdn)
                }

                return acm.NewCertificateValidation(ctx, "certValidation", &acm.CertificateValidationArgs{
                        CertificateArn:        s.cert.Arn,
                        ValidationRecordFqdns: fqdns,
                }, pulumi.Provider(s.usEast1))
        })

i can have it return

CertificateArn

but

ApplyString

converts that

StringOutput

to an

AnyOutput

it appears, which i can’t then use as the input for creating the distribution

I've got an issue where I'm trying to specify

delete_before_replace

but Pulumi seems to be ignoring that advice and still trying to create the resource before deleting the original. In this case it's a Vault Mount that is part of a component resource. I've tried passing the resource options to the entire component, and to just the Mount resource with no luck.

Nevermind, I figured out what my issue was. I had changed the URN prefix for the component resource.

what’s the canonical way to get a reference to an

awsx

object when all you have is the

aws

version, without creating any extraneous resources? For example I have an ecs Cluster (created via

aws

) that I wanted to use as the cluster for a Fargate service created using the

awsx

modules, but when I do a

new awsx.ecs.Cluster

and pass in the existing

cluster

as an argument, Pulumi creates some extra security groups

how do i get the URN for a given resource? I'm trying to use an Alias to move a resource around, but can't seem to get the URN right. I don't see it listed on the pulumi site project.

pulumi config set

seems to sort my yaml file - is there any way to turn that behavior off? I prefer logical grouping for a structured config over alphabetical order. I have a few encrypted secrets which I set through the console while config values are added manually with a logical grouping.

Hi all! I tried to join this Slack yesterday, but couldn't - I am unable to deploy an EventGridSubscription to Azure, anyone have experience with this? Because I couldn't join slack, I created a Github Issue that actually describes my problem here: https://github.com/pulumi/pulumi/issues/6048 The error I get:

azure-nextgen:eventgrid/v20200101preview:EventSubscription (EventSubscription):
    error: Code="InvalidRequest" Message="This event subscription operation is not supported using this API call. Please use the proper API and try again."

Help!

Hey I'm trying to setup an ECS fargate Service with multiple portMappings (listeners) and running into this https://github.com/pulumi/pulumi-aws/issues/1258. I tried figuring out where the problem comes from but I think it's more deeply architectural in that networkListeners only has one NetworkListener per container instead of an array, so it's probably a larger fix. Until such time...has anybody found a successful workaround to have a ECS Service listen on multiple ports via a LB?

Hello everyone! I've had some fun this weekend trying to make some components that can help setting up a simple Wordpress site using Google Cloud Platform and Cloud Run – would love some feedback! https://github.com/cobraz/pulumi-wordpress

I'm using

pulumi/docker

, and can't find a way to skip building a Dockerfile. In my use-case; I have a mono-repo with 10+ Dockerfiles, and most builds they don't have any changes, therefore I would like to skip building them. I've been looking at the documentation, but haven't been able to find a way to introduce that. Wondering if

context

variable is what I should use, or if its even supported?

Hey guys, about the

pulumi import

cli command, i think the docs can be improved a bit. The usage is

pulumi import [type] [name] [id] [flags]

but only the flags are explained, is not said anywhere what

type

means, or how can you know what is the type you are supposed to use for some specific resource or where you can found it. And the others, name, and id, are more intuitive, but still the docs say nothing about them Also i saw that some resoucerces, like aws. now include a section with the proper cli command for import them ,but others like github, don't have it. I was specifically looking to import a github repo, but was unable to find how.

Does pulumi support GCP logging buckets? According to [this GCP documentation](https://cloud.google.com/blog/products/management-tools/cloud-logging-adds-log-buckets-feature) it appears that Cloud Logging Buckets are distinct from Storage Buckets. I see that Pulumi has a

gcp.logging.ProjectBucketConfig

but I don’t see a way to create Logging Buckets - only

gcp.storage.Bucket

Have anyone experience

pulumi stack ls -a

not returning all the stacks? I'm using azure as backend at the moment and running that command only returns two out of many more stacks. This is run inside a docker container (which I can explain why if needed) with the environment variables to login set from the current shell. Running the same command outside docker returns the expected result.