image.png

Is it possible to disable downloading plugins and provide a custom path to them?

It seems if the plugin executable is already found, it doesn't install it into ~/.pulumi/plugins. I wanted a way to prevent automatically installing whichever plugin pulumi can't already find

Can I have multiple versions of the same plugin? If so how can I specify a function or type from a specific version?

has anyone played with Alchemy? the creator is making some fairly inaccurate claims about Pulumi

I started Pulumi with Java. The multi-language support was its best selling point. BTW, Alchemy is a stinking word.

Has anyone compared pulumi-github with the GitHub Provider for terraform? In my experience, the terraform provider is fiddly, incomplete, and not yet ready for automation - at least not for all the GitHub infra I'd like to manage. Is pulumi-github substantially better? (Even though I assume it, like the tf provider, both leverage the same golang API?) OTOH, I've found very few (2?) reports from anyone using IaC to mange GitHub itself, so I wonder if I'm looking at the problem in the wrong way. Like, maybe what I need is an IDP, Backstage scaffolding, or the like. Or maybe I should separate user management from repo management, and use different tools for each? (Which are generally used by different groups of users) Any pointers appreciated.

I'd like to run

pulumi up

with

--yes-if-nothing-dangerous

which, of course, doesn't exist. The idea would be to act like

--yes

but only if there are no deletions or replacements. • are there any plans to add this sort of functionality? • has anyone come up with something that does this short of dumping preview/etc output and glorified

grep

ing it?

Does

pulumi refresh --stack org/project/stack

always require a

Pulumi.yaml

file? We rely on the automation api extensively, so lots of time there isn't a

yaml

file available.

We are currently trying to migrate existing stacks from snowflakes old user type to the new service_user type (https://www.pulumi.com/registry/packages/snowflake/api-docs/serviceuser/). I am struggling to find a decent approach on how to achieve a clean migration. The user that we want to migrate is at the same time utilized in the provider specification for some other objects which can not be deleted and recreated. Does anyone have a proposal on how we should handle this migration?

quick question: how do you add a "Package" in a java automation?

I have included the generated sources for the SDK, but am guessing since in automations I don't have an actual Pulumi.yaml I need to also register the package / make it available, etc

if i am using the automation API and Is there a way to get the current stack data and modify the resources name and all then create a new stack using it

Hi. We're looking into creating an resource in project/stack A that is references with a stack reference in another project/stack B. When we modify the resource in A pulumi decides that it needs to be replaced and the original A is deleted. A is a resource group in azure which means that all resources created in that resource group in B also will get deleted. Is there a way to have Pulumi report that those resources in B will get affected by the changes in A when we update the resource in A?

We're using the pulumi cloud as backend, so I'd assume that it could lookup if the stack reference is in use at least somewhere.

Is it possible to detect (from code) what policy packs are enabled?

🚀 Nuevo episodio de CursaCloud 🎙️ "De YAMLs a código real: ¿es Pulumi el futuro del IaC?" ¿Y si pudieras escribir tu infraestructura con el mismo lenguaje que usás para programar? Sin YAMLs eternos. Sin configuraciones repetidas. Sin limitarte a lo declarativo. 🔗 Escuchalo acá en Spotify: https://open.spotify.com/episode/6S9szWdOr5AvJ2eRNZ1xu6?si=Lu81R6_GSK2R72oqNrKotA

Any plans on a more fully supported redpanda package https://www.pulumi.com/registry/packages/redpanda/?

Hi. Is there something wrong with the pulumi installation? the screenshot is Github Action failing when trying to access https://get.pulumi.com/releases/plugins/pulumi-resource-std-v1.7.2-linux-amd64.tar.gz

Hi, I think I need some clarification/guidance on stack selection. We do have pulumi in CI/CD running for a while now, and are doing pulumi stack select and pulumi up in different commands similar to https://www.pulumi.com/docs/iac/using-pulumi/continuous-delivery/gitlab-ci/ - but somewhere between pulumi 3.154 and 3.171 the stack selection ceased to be persistent. I can reproduce it with local docker containers and stack select then stack ls and can see the stack in ~/.pulumi/workspaces/ Some LLM chat claimed deprecation and using stack parameters all over being the way to go, but references since vanished docs. What is the official situation here? Bug? Deprecation? Am I holding it wrong?

Has anyone got a working example of a github action that uses uv with pulumi?

hello newbie here has anyone got pulumi automation api working with gcp i am using pulumi access token

import os from dotenv import load_dotenv import pulumi from pulumi import automation as auto from pulumi_gcp import compute # Load environment variables from .env load_dotenv(r"C:\Users\USER\OneDrive\Desktop\scaling_considerations\prod\deployment_manager\.env") # 1. Define the Pulumi program def pulumi_program(): # Create a VPC network compute_network = compute.Network("network", auto_create_subnetworks=True) # Allow HTTP and SSH access compute_firewall = compute.Firewall( "firewall", network=compute_network.self_link, allows=[compute.FirewallAllowArgs(protocol="tcp", ports=["22", "80"])], source_ranges=["0.0.0.0/0"], ) # Reserve a static external IP instance_addr = compute.Address("address", region=os.environ["GCP_REGION"]) # Define instance metadata to run a Docker container container_metadata = { "gce-container-declaration": """ spec: containers: - name: hello image: nginxdemos/hello ports: - containerPort: 80 restartPolicy: Always """, "google-logging-enabled": "true", } # Launch the VM with Container-Optimized OS and the container compute_instance = compute.Instance( "instance", machine_type="e2-medium", metadata=container_metadata, boot_disk=compute.InstanceBootDiskArgs( initialize_params=compute.InstanceBootDiskInitializeParamsArgs( image="cos-cloud/cos-stable" ) ), network_interfaces=[ compute.InstanceNetworkInterfaceArgs( network=compute_network.id, access_configs=[ compute.InstanceNetworkInterfaceAccessConfigArgs( nat_ip=instance_addr.address ) ], ) ], zone=os.environ["GCP_ZONE"], service_account=compute.InstanceServiceAccountArgs( scopes=["https://www.googleapis.com/auth/cloud-platform"] ), opts=pulumi.ResourceOptions(depends_on=[compute_firewall]), ) # Export outputs pulumi.export("instanceName", compute_instance.name) pulumi.export("instanceIP", instance_addr.address) # 2. Use the Automation API with env values def main(): stack_name = "dev" project_name = "gcp-vm-docker" # Load environment variables gcp_project = os.environ["GCP_PROJECT"] gcp_region = os.environ["GCP_REGION"] gcp_zone = os.environ["GCP_ZONE"] # Create or select a stack stack = auto.create_or_select_stack(stack_name=stack_name, project_name=project_name, program=pulumi_program,) print("Installing plugins...") stack.workspace.install_plugin("gcp", "v6.0.0") # Adjust version as needed print("Setting config from environment variables...") stack.set_config("gcp:project", auto.ConfigValue(value=gcp_project)) stack.set_config("gcp:region", auto.ConfigValue(value=gcp_region)) stack.set_config("gcp:zone", auto.ConfigValue(value=gcp_zone)) print("Refreshing stack...") stack.refresh(on_output=print) print("Updating stack...") up_result = stack.up(on_output=print) print(f"Instance name: {up_result.outputs['instanceName'].value}") print(f"Instance IP: {up_result.outputs['instanceIP'].value}") if name == "__main__": main()

Hi all, We see a lot of requests to our state backend when doing a

pulumi up

without any changes. Is this expected behavior? I would expect Pulumi to only make a snapshot if something changes, is that not the case? We have found the skip-checkpoints flag, but would like to better understand the underlying issue (if any) before we enable it. Our large >3000 resource stacks are currently taking a very long time (>40min) to

up

even if there are little or no changes and are looking to speed it up.

Is there any way to enforce on a stack or org level what pulumi client version is being used? Essentially want to confirm that users are using the same version, or at least a sufficiently updated one

Can a dynamic provider depend on outputs from another resource?

I ran into an issue where I have a set of resources with dependencies between them, where the graph I want to create is a DAG (ie. no cycles) but because of the way the resources are grouped into ComponentResources, there's no way to set up the dependencies correctly. Is there a way to add a dependency to a resource after creating it?

What is the recommended way to promote code from one environment to another, for example from dev to stage then to prod? I can use the pulumi project and stacks constructs but im not sure if that gives me the controls that i need. For example if i create an AWS EKS cluster in a project and launched it to all stacks (dev, stage, prod). Then i need to make a change, something that is not as easy as a parameter that can be in the stack environment config, something like changing networking cni. This will be a bunch of various code changes and it could take a while to roll out with testing along the way in the various environments (dev, stage, prod). Once i have made this change in dev and it is working for me, i would like to open a PR to get it reviewed, merged and then versioned so that the next environment can use it. I would like that the project or stack can reference a github URL that can ref a tag/commit. I am coming from the Terraform world where you can do this via:

module "consul" {
  source = "github.com/hashicorp/example?ref=v1.2.0"
}

Is there something similar in pulumi? Does this make sense to do or am i trying to follow too closely to what i have done in Terraform? I have found Pulumi Components: https://www.pulumi.com/docs/iac/using-pulumi/extending-pulumi/build-a-component/#sharing-via-git. This gets me most of the way there but i still have to run a

pulumi package add <repo_url>@<release-version>

Which means i have to run this command and the CI/CD pipeline has to run it as well. This would mean adding a script that runs this command with the version? So the person updating it would have to update the version here? Not the worst thing but the workflow seems to be just a little off as in it is making the user look into another file that is not a Pulumi file to update a version number. Other notes: • i am only using python • I dont want to publish these publicly. Not that there are any secrets or proprietary things in it. For now, we dont want to spend time to make it publicly consumable.

#C84L4E3N1 I am currently only working on ABC guest posting. If you have opportunities for that, please let me know. I am interested in link exchanges with SaaS sites. I need sites with stats, DR 50+, and Traffic: 5K+. Must be from Tier One countries. Likewise, the UK. US, CA, AU