are there any examples of mocking a stackreference?

how can we specify the

kubecontext

for a Helm chart resource?

Talk on that compared

pulumi

with other tools in

kubecon

https://kccncna20.sched.com/event/ek9o/five-hundred-twenty-five-thousand-six-hundred-k8s-clis-phillip-wittrock-gabbi-fisher-apple

anything like pulumi2tf available? use case: my org is currently bound by contract to supply tf scripts to our customers for infra on which our software suite is installed to even consider migrating to pulumi, we must first achieve the intermediate goal of generating same tf from equivalent pulumi

pardon if this is not an appropriate channel for this, but I am looking to bring on a devops engineer to help me promote and build a great devops culture. We are a small and young tech company that is really well funded. I and the dev team want Pulumi to be at the heart of our IaC! If interested, DM me for info. (not sure I’m allowed to post it here)

what's the recommended way to run shell commands with pulumi after a resource is created?

If I have a resource that contains a secret, then

pulumi update

will only show the details as

secret

. For example, on an AWS Task Definition:

~ containerDefinitions: "[secret]" => "[secret]"

I really want to see what's about to change on that

containerDefinition

. Is there a way to show those secret values?

is it possible to pass values from an API as configuration values to the Pulumi script? I don’t want to set the config values in the CLI mode, I am looking for a fully automated way without using CLI and the Pulumi script gets the value from another source and use it to do the provisioning.

Hello folks, When i am trying to create Ingress/Egress attribute of the SecurityGroup, it throws a following error

ingress=[aws.ec2.SecurityGroupIngressArgs(
    AttributeError: module 'pulumi_aws.ec2' has no attribute 'SecurityGroupIngressArgs'

The code was referred from an Pulumi docs example.

security_group = aws.ec2.SecurityGroup( # type: ignore
    "secgrp",
    description="Splunk Hybrid cloud security group",
    vpc_id=vpcid,
    ingress=[aws.ec2.SecurityGroupIngressArgs(
        description="TLS from VPC",
        from_port=443,
        to_port=443,
        protocol="tcp",
        cidr_blocks=["0.0.0.0/0"],
    )],
    egress=[aws.ec2.SecurityGroupEgressArgs(
        from_port=0,
        to_port=0,
        protocol="-1",
        cidr_blocks=["0.0.0.0/0"],
    )],
)

Any suggestions?

Ran into an issue today, looks like when creating a kubernetes secret with Pulumi you have no choice but to accept autonaming of the secret. https://github.com/pulumi/pulumi/issues/5790

How do I create a windows node AKS? I am trying to build mixed node cluster and I cannot find any examples on how to achieve this.

Hi! I am playing with

random.RandomPassword

and I'm wondering why it's

result

output is not marked as secret and I have to do:

const password = new random.RandomPassword('myPassword', {...}, { additionalSecretOutputs: ['result'] })

Shouldn't it be marked as sensitive by default?

Hello, My stack is really simple with a single component :

export const loadBalancer = new lb.ApplicationLoadBalancer('alb');

No param is given so it defaults to default vpc, which as subnets in two availability zones, and still I see :

error creating application Load Balancer: ValidationError: At least two subnets in two different Availability Zones must be specified

Any idea on what is happening ? Thx

Hello everyone! I have a general question, does Pulumi have a way to apply tags to all resources in a stack?

Is there a concept of a

local

stack? What I mean is, before pushing to the

development

branch where my github actions run

pulumi up

I want to test out if everything works one step before, where a stack gets created automatically, resources are created in the cloud, things get deployed, and when no errors are returned by pulumi and I am satisfied with my deployment, I would then merge the branch to the

development

branch and the

local

resources and the stack are destroyed? Sometimes when I make some changes,

pulumi preview

doesn't return any errors, because I have defined everything correctly, but in reality my setup i.e port mapping to my ECS instance has not been set up properly, so I want to go back, modify my config and try again. Is that something possible? I have a feeling that's what the automation API is capable of.

Is there a way to gather timings of different resources. I have something which is taking some time during preview and wondered if there is an easy way to track it down

Hey everyone, is there a way to access the current stack name in code? I'm trying to work around SQS FIFO queue name issue (#967). I'm happy to provide my own name, but I'd like it to be unique inside the stack

is there a way to force pulumi destroy to delete s3 bucket objects?

error: deleting urn:pulumi:dev::crytic::aws:s3/bucket:Bucket::my-bucket: 1 error occurred:
    	* error deleting S3 Bucket (my-bucket-467da8e): BucketNotEmpty: The bucket you tried to delete is not empty

I'm currently deleting an RDS instance. The instance is gone, but the optiongroup won't delete, because it's referred to by all the auto-backups that RDS has made. It'll timeout soon and I'll have to go through the manually-delete/export stack/update stack/import stack rigmarole to get it sorted. Is there an alternative? Maybe something like: * Pulumi could discover unmanaged resources that would prevent a deletion (e.g. BucketObjects, backups, etc.), * mark some types of resources as "can be blocked on deletion" and given them a shorter timeout, a more-specific failed-to-delete error message and don't mark them as deletion-in-progress? * A

pulumi undelete

command that marks the resources as not-deleted instead of deletion-in-progress (so that I can go delete the unmanaged resources then run

pulumi up

again, cleanly).

Is there any way to not store some data in the state? I've found that when deploying cert-manager helm chart using Pulumi the state file grows from around 200Kb, to 8Mb, due to the amount of CRD's in the chart. Is there anyway to get these deployed without inflating the state and making everything slow

are there any plans for a rook+ceph provider?

Doc improvement suggestion: the Pulumi type could be added to all resource doc pages. It's used in a few things like

pulumi import

and stack transformations, but to find a type I (think I) have to open node_modules/@pulumi/<provider>/<somedir>/<resource>.js and look down the bottom of the file.

hey guys, my team is trying to migrate from heroku to aws ecs. On heroku we used to have 'rollback' features. Is there any way to do similar thing in pulumi?

I'm working with

pulumi import

. I've attempted to import just a few SNS topic policies (I'd successfully imported the topics already). I'm using a json file, and I've set a nameTable with the conveniently-but-confusingly named this and provider properties: I chose this because the generated code was to go in the constructor of a ComponentResources, and the parent of the topics would be

this

. Short cut! Importing worked but no code was generated. The state was updated and looks like I want it to. The parent and provider properties of the new resources have the correct values (values from the nameTable properties). This error was output:

error: no name for provider urnpulumiMyStack:MyProjectNameawsMyClassmyId

What might have gone wrong?

i have to say, i’m not very familiar with devops tooling (i’ve been meaning to look into terraform) and I decided to use pulumi and it’s amazing, big fan. It saved me and my team so much time.

I really like the graph projection, both in the ui and the terminal. But I haven’t yet found documentation on how to easily create “parent aggregators”, something like a dummy component. Is there such a default component? So far I’ve been doing this (python), but I have a feeling it is not best practice:

from pulumi.resource import ComponentResource

class AggregatorComponent(ComponentResource):
    def __init__(self, *args, **kwargs) -> None:
        super(AggregatorComponent, self).__init__(t="aggregator", *args, **kwargs)

and using it like so

project_aggregator = AggregatorComponent(name=project_id)
opts = ResourceOptions(parent=dataset_aggregator)
DemographyDataset(opts=opts)
FunctionsDataset(opts=opts)
...

Is there a better way to do this than my homemade

AggregatorComponent

?

Is there a way to create a non-secret output from a secret output?

How does one use async block inside a ComponentResource? I know the pattern of encapsulating my regular pulumi TS code in a big async block. But now I have a ComponentResource class that does some async gets on resources. Do I have to use .then chaining?

export class CustomComponent extends pulumi.ComponentResource {
  constructor(name, args, opts) {
    super('pkg:CustomComponent', name, {}, opts)
    
    const computeDefaultServiceAccount = await gcp.compute.getDefaultServiceAccount({project: projectId});
    gcp.projects.IAMBinding('blah', { role: '', project: projectId, members: [`serviceAcccount:${computeDefaultServiceAccount.email}`]}
  }
}

Hi! I try to run pulumi import awseks/clusterCluster <name> <id> and I receive this error

error: aws:eks/cluster:Cluster resource 'resource' has a problem: Computed attribute cannot be set

How do I access the auto named resource inside the resource? I need to basically access

self.name

but I want the generated name. I am using python.