Apologies. Not sure why it defaulted to sending this to channel again, only did that with the bump on purpose. 🐑 ish grin.

Devspace/Skaffold workflows with pulumi. https://github.com/pulumi/pulumi-kubernetes/issues/1620 Anyone using a dev workflow tool for easier reloading locally of Go app changes, while still leveraging pulumi? The automatic port forwarding and such seem great and possible to replace my clunky docker compose setup now since I need more external services from cluster anyway.

I'm trying to create a suspended job to run later with pulumi but for some reason its creating pods and pulumi is waiting for the execution of the job - is this normal?

I’m a little unclear as to how Pulumi Kubernetes Job works - does it register success after the Job is complete or after the Job is created? The docs mention “ready” status, but I’m not sure if “ready” means Job complete or Job created (i.e. started)?

Anyone who can shed some light on this would be greatly appreciated

what's the difference between • https://www.pulumi.com/registry/packages/eks/api-docs/managednodegroup/ • https://www.pulumi.com/registry/packages/eks/api-docs/nodegroup/ it seems like they both create managed node groups, based on what I see from the AWS EKS UI?

Hello! I couldn’t import the already deployed helm release from my k8s cluster. I wonder its intended behavior. Could anyone help with this issue?

Hello guys, I’m using the

kuberenetes.helm.v3.Release

component, and wish to know if there’s a way to instruct the pulumi API to use a specific revision number. Or perhaps is there a way to perform a helm rollback instead?

Is there a quick fix for persistent volumes not deleting fast enough? Specifically fromt he nitnami postgresql chart

-  kubernetes:core/v1:PersistentVolumeClaim hasura-46b9 deleting error: 'hasura-46b9-xfv0a937' timed out waiting to be Ready
-  kubernetes:core/v1:PersistentVolumeClaim hasura-46b9 **deleting failed** error: 'hasura-46b9-xfv0a937' timed out waiting to be Ready

👋 sup yall. im a little stumped by some k8s / helm + pulumi behavior, hoping someone might be able to shed light i’ve got a: • custom ConfigMap resource, with a YAML string as the

data=

input (in a key) • a helm chart resource, that takes ☝️ (

ConfigMap.id

) output in the

values=

input object i’m hoping to get one of the helm chart children resources (eg. a Pod) to update whenever the ConfigMap data value gets updated in our workflow - basically, hoping that they both get updated together based on the inputs/outputs docs, I’m instantiating things in this order: 1. instantiating ConfigMap resource 2. assigning the output via apply()

some_object["child_"resource"]["config_key"] = ConfigMap.id.apply(lambda v: v)

3. instantiating helm chart resource, and setting

value=some_object

…but when the ConfigMap data gets modified, I see ALL the helm resources just get

deleted

(not even replaced). but when the ConfigMap data is unchanged, everything is gravy. What might be the causing this? i’m getting lost in the various layers of abstraction

I'm trying to deploy a simple helm chart with Python and

pulumi_kubernetes.helm.v3

but I'm getting an error

failed to pull chart...version...not found

. I can install the chart with Helm v3 fine. What could be my problem?

Does anybody know how I can update the

default

namespace with labels and annotations using Pulumi? When I'm using https://www.pulumi.com/registry/packages/kubernetes/api-docs/yaml/configfile/ I'm getting the following error:

error: update failed

  kubernetes:core/v1:Namespace (default):
    error: resource default was not successfully created by the Kubernetes API server : namespaces "default" already exists

on deploying a recent change to AWS Load Balancer Controller via helm, pulumi seems to be having issues with naming on one of the sub-resources

IngressClassParams

.... the first time i run pulumi the pulumi logical-name of this resource is

kube-system/alb-ingress-class-params

but any subsequent run tries to replace it with another instance with the name

alb-ingress-class-params

(ie without the namespace prefix as part of the logical name). ... thoughts ??? ..... more details in thread

Hi! I have k8s deployed on DigitalOcean via Pulumi, and I also deploy my Helm releases onto said cluster using Pulumi. If I try to update an existing Helm release, or create a new one, I get the following when I run `Pulumi up`:

error: can't create Helm Release with unreachable cluster: unable to load schema information from the API server: the server has asked for the client to provide credentials

I have my digital ocean access token set up, and I initially created the Helm release with Pulumi, so I'm not clear as to why it suddenly isn't working. Any guidance or help would be massively appreciated!

Hi all, Has anyone managed to use eks entirely from credentials stored in the stack config? I’ve got it working on up by actually retrieving the token and shoving it in a kubeconfig but destroy doesn’t appear to re-retrieve the token, just relies on the one stored in the state.

Hi all, At the moment I have a cronjob that makes a snapshot every x minutes of a given pvc in a namespace. It does this by using a new image using the Pulumi cli image as a base image and adds a small Pulumi project that makes a new snapshot and deletes the previous one every-time it runs. Now this is all working fine, but there is one small issue. I need the value

Snapshot Handle

from the volumesnapshotcontent resource that is generated automatically when making a new snapshot. Because it's generated automatically it is not a pulumi resource, so as far as I know I can't get that value using pulumi. I wanted to use pulumi for it so I can save it as a stackreference and use it in different places in my infrastructure. I tried using query-kubernetes, because I can get the name of this volumesnapshotcontent from the volumesnapshot itself, but this package does not support crd's. Anyone an idea on how I could achieve this? (I thought of using a kubectl container to get it, but I would prefer to keep it all pulumi)

Hi everybody, Question: What’s the best way to rename a helm release without downtime? Context: I made a mistake naming a

helm release

and wanted to renamed it. Ideally, I wish to prevent pulumi/helm from creating a replacement, so I tried to do so by using the

aliases

flag on the

ComponentResourceOptions

. However, after aliasing and changing both the name of the pulumi resource and the

helm.Release

component

name

field, I noted that no changes are being applied.

Hi there! 👋 Has anyone found way to install Flux v2 with Pulumi?

I broke out my various components in Go into seperate packages. I now am trying to eliminate any of the functions that are overriding the names. I thought i was using the autonaming, but deployment for example took my name and didn't append anything extra. So now, I want to know 2 main things (posting as seperate blocks so reply in threads will work for each...

1️⃣ Not Overriding the Autonaming, but still keeping it easy to share _edit_: removed main content and placed in How do you approach passing around ID's for resources without it becoming a hot mess? #9205 This is more involved so better to capture any ideas there for others. Welcome any insight/thoughts.

2️⃣ multicluster deployment Just modified my plan to do:

---
clusters:
  - name: cluster1
    kubeconfig:
        secure: .....
  - name: cluster2
    kubeconfig:
        secure: .....

Now I just added in main a new config object and wrapped the entire deployment setup I had in a for each. Boom. Loop over clusters and go. Due to pulumi's routinues and all the magic going on I figured I'd do a dumb old loop instead of trying concurrent deployments, but looking for a sanity check.

pulumi.Run(func(ctx *pulumi.Context) error {
  for _, cluster := range clusterList {
	cfgMap, err := configmap.ConfigMap(
				ctx,
				renderProvider,
				&configData,
				&appConfig,
				&sharedConfig,
			)
     err = ingress.NetworkingIngress(ctx, renderProvider, &configData, &sharedConfig)
     err = service.Service(ctx, renderProvider, &configData, &sharedConfig)
     //     etc....
	_ = <http://ctx.Log.Info|ctx.Log.Info>(fmt.Sprintf("finished processing cluster: %s", cluster.Name), &pulumi.LogArgs{})
    }
})

Make sense to do the same deployment this way on 2 clusters without concurrency or any other way you'd say is better? Multiple pulumi.Run or anything else you'd recommend? I love my experience overall, but it's a lot more complex initially than plain yaml, so I'm double check my approach here makes sense and is reasonable. Much appreciated!

kubernetes.networking.v1.Ingress

appears to expect a LB address to populate in status before accepting the resource as being healthy? I'm not seeing an opt-out for this which is causing difficulty behind noteport / hostport type ingress controllers. Am I missing something, or should I be trying to accomplish this another way?

Hello Community! I am trying to use the imported

kubeconfig

from another project to create a new provider on the new projct based on that file using the

ApplyT

golang method to convert the

AnyOutput

data type to a

pulumi.ProviderResource

type. Can someone explain to me how can I do it? A snippet of golang code is below.

// Create stack reference to prvious projct named  "ccmx/eks/dev"
stack_output, err := pulumi.NewStackReference(ctx, "ccmx/eks/dev", nil)
		if err != nil {
			log.Fatalf("error creating stack reference: %v", err)
		}
// Import kubeconfig from the stack "ccmx/eks/dev"
kubeconfig := stack_output.GetOutput(pulumi.String("kubeconfig"))

// Create Kuberntes provider
K8sProvider, err := kubernetes.NewProvider(ctx, "eks-provider", &kubernetes.ProviderArgs{
			//Kubeconfig: pulumi.String(kubeconfig),
			Kubeconfig: kubeconfig.ApplyT(func(s string) string { return s }).(pulumi.Provider),
		})

		if err != nil {
			log.Fatalf("error creating K8s provider: %v", err)
		}

Is there any good information or insight someone can give on the difference between helm/v3 charts vs releases? They seem to do the same thing?

Is it possible to tell pulumi about dependencies within resources in yaml files using NewConfigGroup? For example, the code below adds three yaml files of resources. The last file

eck_filebeats.yaml

has a resource named

filebeat

that should not be created until a specific service created by the

eck_operator.yaml

file has an endpoint.

Also, how can I tell pulumi to expect that the

eck_operator.yaml

file will deploy an operator pod that will, in turn, deploy a service outside of Pulumi, and that this service should be imported into Pulumi so it can be dependedOn by other Pulumi-managed resources downstream?

Hi everyone! How do you provision node-groups in EKS? I’ve tried to do same as noted in EKS Crosswalk, but have infinity preview state until I’ll remove NodeGroup opts (with provider settings). But then I’ll have problem with broken assume_role_policy as showed on screen (again, I’m still using example from link above). Then I’ll use my one custom assume_role_policy (see second screen), but after it provides EKS I’m getting error with Stack (third screen) I’m really need custom node-groups… How can I do that? Thanks! :)

Does pulumi have an approach to handle the k8s operator pattern? When operators are used for an app, they create resources outside of pulumi. How does pulumi deal with this? Is there a operator "monitor" mode where pulumi can pick up these dependencies dynamically from the operator's interaction with the k8s API?