When using the Automation API, how can I override the default providers? Background: I am writing end-to-end tests. First, I deploy a GKE stack. It provides a kubeconfig in its outputs and I would like to reuse it to deploy stuff on the cluster. I know I can pass the kubeconfig to the 2nd pulumi program as a config value, but that would require boilerplate code I'd like to avoid (because there are many pulumi programs i want to programmatically deploy in my tests).

question about using Helm - what version of Helm is used by the cli in the Action? Is this configurable? if not, how can we ensure we know what version is used so that I can keep my local environment in sync with that version of Helm based on which version of Pulumi I’m using?

Hello, I am building an HTTP API using the Automation API for managing dynamic Pulumi stacks based on HTTP request parameters. The API provides actions like

create

,

update

, and

delete

for managing the stacks. My question is, what is the correct workflow for updating all stacks when I need to reconcile a new version of the Pulumi code for existing stacks? Do I need to get all stack identifiers and run an update on each one individually?

I’m using Typescript. when running the following lines I can see that it’s takes a lot of time much more when I run it using the cli

const prevOpts: PreviewOptions = {}
    const changes = await stack.preview(prevOpts)
    console.log(
      `preview summary: \n ${JSON.stringify(changes.changeSummary, null, 4)}`
    )

since the preview works in front of the existing state file which is located in s3 it’s very time consuming

Hi, Is there any difference between

stack.up

of automation api to

pulumi up

of CLI ? I can see that when running using the automation api it takes much more time moreover when running the automation api script twice it does the update all over again - how come ? please advice.

Random question for the group: how did you find out about automation api?

where can I add more verbose logging to my `` call?

does setting

on_output

conflict with other logging properties like

log_verbosity

?

Hello, while I’m trying to execute a pulumi autiomation within my microservice (running in kubernetes / docker container) I’m continuing to have the same error of

Error: Command failed with EACCES: pulumi version --non-interactive

. I’ve installed pulumi in the docker image and linked the PATH, but nothing changed. Any idea about permissions or any other requirements in docker image?

@lemon-agent-27707 Do you have time for a conversation on the new deployments API? We’ve built quite a bit of tooling around the automation API, and I’d be interested to hear what it could do for us

Heya, using auto.UpsertStackLocalSource() how can I specify the passphrase?

Hi, When running

pulumi.up

using automation api does the updates supposed to being executed in a raw or parallel ? This is my code and it looks like it’s running sequential and not parallel

const opts: UpOptions = {
      onEvent: logEventUpdateOutput,
      parallel: 2147483647,
    }
    const upRes: UpResult = await stack.up(opts)

please advise, Is it something that I can update in order to faster deployment ? to make it run in parallel ?

How to make the stdout with color?

stdoutStreamer := optup.ProgressStreams(os.Stdout)

	// run the update to deploy our fargate web service
	res, err := s.Up(ctx, stdoutStreamer)
	if err != nil {
		fmt.Printf("Failed to update stack: %v\n\n", err)
		os.Exit(1)
	}

Anyone have any thoughts on this problem? It may be my container skills and nothing to do with Pulumi, but we are using the automation API running in a Linux container and when we get to running the " await LocalWorkspace.CreateOrSelectStackAsync(stackArgs)" command we get this error: " An error occurred trying to start process 'pulumi' with working directory '/tmp/automation-3g4geg4m.bps'. No such file or directory" I've checked and that folder definitely exists, it is owned by root and the container is running as root, so I can't see why it can't see it. Does the automation API run as a different user to the container? Any other ideas?

Hi, I’m trying to force pulumi to use a specific plugin version (It’s using aws

5.25.0

, and I’m trying to use

4.0.0

), but it’s still installing the default plugin version during the

up

command What I tried - • Change `package.json`/

yarn.lock

files to the specific version:

"dependencies": {
    "@pulumi/pulumi": "^3.50.2",
    "@pulumi/aws": "4.0.0",
    "@pulumi/awsx": "^0.40.0"
  }

• Remove all the installed plugin versions but the wanted version.

pulumi plugin ls
NAME        KIND      VERSION  SIZE    INSTALLED      LAST USED
aws         resource  4.0.0    255 MB  5 seconds ago  5 seconds ago

• Update the

yaml

file with the

plugins

option, as here - https://github.com/pulumi/pulumi-hugo/blob/e59734a63d61722cf18720c6b73bd4af72ca8c38/themes/default/content/docs/reference/pulumi-yaml.md#plug[…]ions

plugins:
  providers:
    name: aws
    version: 4.0.0

I also found this issue, but I couldn’t understand how to do that - https://github.com/pulumi/pulumi/issues/7702 Can anyone help? Thanks

One thing is unclear about Automation API documentation:

The Pulumi Automation API is a programmatic interface for running Pulumi programs without the Pulumi CLI. Conceptually, this can be thought of as encapsulating the functionality of the CLI (pulumi up, pulumi preview, pulumi destroy, pulumi stack init, etc.) but with more flexibility. It is a strongly typed and safe way to use Pulumi in embedded contexts such as web servers without having to shell out to a CLI.
...
Note: Automation API requires the Pulumi CLI to be installed and available in your PATH environment variable.

Do I need the cli binary to be accessible? I'm trying to see if I can embed pulumi in a service without any disk access.

in other words, is the golang sdk self-contained.

Seems I have my answer after looking at github's issues: https://github.com/pulumi/pulumi/issues/7219

Hello, I am trying to create multiple dns records in cloudflare using the automation api. And as such have a

List<string>

of ips that should be mapped to a single domain, this is entirely supported by Cloudflare by just creating multiple records. However as the Ip's comes in as an

Input<List<string>>

because I first have to provision the necessary infrastructure I have no other choice but to use an

apply

to get the underlying Ip's to be able to create the records. However, as per the documentation for apply (https://www.pulumi.com/docs/intro/concepts/inputs-outputs/#apply) it specifically recommends against doing this. How would I go about creating a list of DNS records based on a variable amount of Ip's when the Cloudflare record type does not accept a list of Ip's but only a

Input<string>

(https://www.pulumi.com/registry/packages/cloudflare/api-docs/record/#properties)? My code looks like this:

public ClusterDnsComponent(string name, Input<List<string>> externalIps, DeploymentRegion region,
        ComponentResourceOptions? options = null)
        : base("dgz:cloudflare:clusterdns", name, options)
    {
        var opts = new CustomResourceOptions()
        {
            Parent = this
        };

        var dnsNames = externalIps.Apply(e => e.Select((externalIp, index) =>
            {
                var record = new Cloudflare.Record($"cluster-dns-{index}", new Cloudflare.RecordArgs()
                {
                    ZoneId = "",
                    Name = NamingHelper.GetClusterDnsName(region),
                    Value = externalIp,
                    Type = Constants.Domain.DsnTypes.ARecord,
                }, opts);
                return record.Name;
            })
        );

        DnsNames = dnsNames.Apply(o => Output.All(o).Apply(s => s));
        RegisterOutputs();
    }

Hello, community! First of all, let me express my appreciation for this beautiful community! I have a question: if while using the automation API my stack will be locked, what can I do (using the automation API) to unlock the stack and make it possible to perform a

stack.up

again? Thanks in advance!

Hi, When using automation API in NodeJS via Typescript, Is it possible to run 2 stacks in parallel?

Hello, community! I’m using the automation API to register a GCP Service Account, I’m having this error (the same stack deployed in pipelines works without any impediment)

<ref *1> Error: failed to register new resource sa-thing-dev [gcp:serviceAccount/account:Account]: Resource monitor is terminating

Any suggestion/idea?

Hello there! Just a newbie . Want to get(FastApi) a list os my stacks(backend in GCP). Any idea?

Hey everyone, I want to create a def(python) to get the value from a created stack to access them in a fastApi webservice. Any any idea?

Hello Guys! I am new here. Nice to e-meet you 👋 I have an issue with creating api key for managing cluster in confluent cloud. I use the latest version of confluentcloud pulumi provider. I have confluent cloud api key created. I want to create an api key for cluster management (it will be used for creating topics and other resources, not accessing data). I tried two approaches. 1) Role binding:

const envManagementAccount = new confluentcloud.ServiceAccount(
...
);

const apiKey = new confluentcloud.ApiKey(`...`, {

  owner: {
    id: envManagementAccount.id,
    apiVersion: envManagementAccount.apiVersion,
    kind: envManagementAccount.kind,
  },
  managedResource: {
    id: clusterConfig.cluster.id,
    apiVersion: clusterConfig.cluster.apiVersion,
    kind: clusterConfig.cluster.kind,
    environment: {
      id: clusterConfig.environment.id,
    },
  },
});

new confluentcloud.RoleBinding(
  `...`,
  {
    principal: `User:${envManagementAccount.id}`,
    roleName: 'CloudClusterAdmin',
    crnPattern: clusterConfig.cluster.rbacCrn,
  },
  { dependsOn: [apiKey] }
);

2) Adding ACLs:

const envManagementAccount = new confluentcloud.ServiceAccount(
...
);

const apiKey = new confluentcloud.ApiKey(`...`, {
  ...
  owner: {
    id: envManagementAccount.id,
    apiVersion: envManagementAccount.apiVersion,
    kind: envManagementAccount.kind,
  },
  managedResource: {
    id: clusterConfig.cluster.id,
    apiVersion: clusterConfig.cluster.apiVersion,
    kind: clusterConfig.cluster.kind,
    environment: {
      id: clusterConfig.environment.id,
    },
  },
});

envManagementAccount.id.apply(
  (saId) =>
    new confluentcloud.KafkaAcl(`...`, {
      kafkaCluster: {
        id: clusterConfig.cluster.id,
      },
      resourceType: 'CLUSTER',
      resourceName: 'kafka-cluster',
      patternType: 'LITERAL',
      principal: `User:${saId}`,
      host: '*',
      operation: 'ALL',
      permission: 'ALLOW',
      restEndpoint: clusterConfig.cluster.restEndpoint,
      credentials: {
        key: cloudApiKey,
        secret: cloudApiSecret,
      },
    })
);

In both cases, I get authorization issue: 1)

error: 1 error occurred: * error creating Role Binding: 403 Forbidden: Forbidden Access

2)

error: 1 error occurred: * error creating Kafka ACLs: 401 Unauthorized: Unauthorized

What’s the proper way to fix it assuming the only key created manually is confluent cloud api key?

^^ please ignore, the only issue was the way of retrieving service account it, the following fixes it:

envManagementAccount.id.apply(
    (envMgmtId) =>
      new confluentcloud.RoleBinding(
        `...`,
        {
          principal: `User:${envMgmtId}`,
          roleName: 'CloudClusterAdmin',
          crnPattern: clusterConfig.cluster.rbacCrn,
        },
      )
  );

Hello! I’m looking for a Node devops to setup Pulumi for use as a PaaS service. Feel freem to msg me if you have experience 🙂

Is there an equivalent

pulumi plugin install <plugin> --file <file>

for an inline application? I'm trying to prevent from having to make calls over the network when installing the AWS plugin via

workspace.installPlugin('aws', 'v4.0.0')

. Ideally, I'd have the binary already available in my docker container and point to it.

Hi, Just for visibility... I found what appears to be a bug in the typescript automatiion api when providing an

auth

section that supplies a

username

and

password

. I've raised it here: https://github.com/pulumi/pulumi/issues/12264