https://www.pulumi.com/docs/reference/pkg/aws/apigateway/integration/ The tabs for selecting language are not working here

is there any project at that time that could used kubernetes as backend of pulumi?

We are getting an error creating a privateLink across subscriptions. (typescript) azure😛rivatelink:Endpoint (sql-link): September 25th 2020 13:49:32 Info error: Preview failed: Error building AzureRM Client: Azure CLI Authorization Profile was not found. Please ensure the Azure CLI is installed and then log-in with

az login

. This is an octopus deploy job using a service principal. The job uses the variables

$env:ARM_CLIENT_ID = $AzAccountClient
$env:ARM_CLIENT_SECRET = $AzAccountPassword
$env:ARM_TENANT_ID = $AzAccountTenantId
$env:ARM_SUBSCRIPTION_ID = $AzAccountSubscriptionNumber

We have dozens of deployments per week across 20 stacks that has been using this process. Its stable. We are trying something new, creating privatelinks from subscription A to a vnet in subscription B. Since most of the code executes in subscription A, the env variable ARM_SUBSCRIPTION_ID is set to "A". However, when we get to the privatelink, we build a new provider and pass into the service principal the credentials, tenant, secret and subscription B. But when we use this new provider in the private link, we still get the same error. We know we have to pass in subscription B to privatelink. Older versions of the pulumi azure provider would default the credenatials, secret and tenant, but the newer version seems to require them. @pulumi 2.1 @pulumi/azure 3.20.1

//Provider Code
const Sub = new vznaz.azure.Provider("subscription B", { //TODO
  subscriptionId: "subscription B",
  metadataHost: '<http://management.azure.com|management.azure.com>',
  clientId: process.env["ARM_CLIENT_ID"],
  clientSecret: process.env["ARM_CLIENT_SECRET"],
  tenantId: vznaz.AADTenantId,
});
vznpulumi.pulumi.ProviderResource.register(Sub).then(p => p);




//PrivateLink Code that is failing with above error message
const privateLinkTags = TAGS;

const PrivateLinkSubnet = vznpulumi.pulumi.output(vznaz.azure.network.getSubnet({
    resourceGroupName: 'vznz-net_rg',
    virtualNetworkName: 'vznz-net_vnet',
    name: 'private-link-hub01',
}, {
    provider: Sub,
    async: true,
}));

//sqlServer
const SqlPrivateLink = new vznaz.azure.privatelink.Endpoint('sql-link', { //was vznaz.azure.privatelink.Endpoint but was getting an error below that privateIpAddress didn't exist on SqlPrivateLink.privateServiceConnection.privateIpAddress
    resourceGroupName: '_prod-rg-01',
    privateServiceConnection: {
        privateConnectionResourceId: sqlServer.id,
        isManualConnection: false,
        name: sqlServer.name,
        subresourceNames: [ 'sqlServer' ],
    },
    subnetId: PrivateLinkSubnet.id,
    name: sqlServer.name,
    tags: privateLinkTags,
},{
    provider: Sub,
});

We are stumped. Any help would be greatly appreciated.

Hi guys, a quick q. I use microk8s to run kubernetes on a linux Hyper-V VM. Would Pulumi still work with this?

I'm running into a very unusual issue that started happening recently (not sure what triggered it)

pulumi up

always updates lambdas (~code). Then:

pulumi refresh

always shows those same lambdas needing update. I let it update. Refreshing / Previewing says no changes need to happen. I then change the memory size of one of the lambdas, but it triggers all Lambdas to update (~code), and one of them ~code + memorySize. This happened recently - I'm using WSL2, I've diff-ed stack exports, and file sizes before and after - and I see nothing that stands out. Any thoughts on what may be the issue or where I should look next?

(also if github is the proper way to post this, happy to post it there)

Hi All, what is the latest best practices on github actions workflows to prevent multiple `pulumi up`s happening on pushes? I see this is used for the docs, https://github.com/pulumi/docs/blob/master/.github/workflows/build-and-deploy.yml is turnstyle the way to go?

Hi there, I'm having trouble to do a Pulumi Lambda to RDS , is there any good example I can look at? I get errno":"ENOTFOUND","syscall":"getaddrinfo from the lambda

Hi all, nice to e-meet you. I have a challenge of rolling out updates to k8s node pools without incurring downtime. Currently, pulumi destroys the old node pool and creates a new one in sequence. My team would prefer if the new node pool was created, pods were allowed time to migrate to the new node pool, and then the old node pool was removed. Does something like this exist in pulumi already? We are a golang shop. Also, are there docs on how to write plugins? I first thought my challenge would be best solved by a plugin but I am not readily finding docs on how to create one. I have found is this ...

To understand how to write and distribute your own plugins, please consult the relevant documentation.

... here https://www.pulumi.com/docs/reference/cli/pulumi_plugin/ but I don't know where " the relevant documentation" is. Many thanks in advance.

Is it possible to unit test a TypeScript pulumi program when using ComponentResources? I can only see resources which don't have a parent set?

I have a weird issue where I get a typescript type error on circleci and no type error when running locally. Same command and all package version, including ts-node are the same. Running

tsc

locally also detects the type issue. Does anybody have any ideas what could contribute to typeissues being ignored by the pulumi runtime locally?

If someone has deleted or changed a resource manually, will pulumi restore it on the next

pulumi up

or will pulumi just check its own state when deploying and not seeing the change? If yes, does that apply to all types of resources?

Does Pulumi have any support for knative ?

Good morning, btw

Another question, I see AWS has elasticache (redis) global support now, I presumed that was multi regional multi az elasticache through the pulumi api,but when i try to stand a redis cluster with 3 nodes for 3 AZs making it multi regional

I get the error that redis is only supported with one node, not more than one

if my k8s deployments are targeting a docker image with the :latest tag and the image is updated.. will "pulumi up -r -y" pull the new image to the pods? and if not is there a command that i could use to do this?

Hello all! I've got a question about reading in stack reference outputs in c#. Is there a way to read in an array of strings? I've stored the strings as an Output<ImmutableArray<string>>

I've got 3 Pulumi projects that work togther, so pre-requisites, infra and application. I've got a stack config in pre-requisites that has a number of values, which are going to be used in all projects, are the only options here either to manually declare options in each project, or to pass them as stack outputs from the pre-requisites? There are quite a lot of values to pass, plus some that are not of type Output<T>.

Another stack related question, in some examples with multiple projects the same stack name is used in each project, making it easier to select which to use. However if I try this I get "error: stack <name> already exists", so how is this possible? I am currently using the local state backend for testing if that makes a difference here

Hi all, I've just started playing around with Pulumi for our organization. I'm using the Go SDK and was wondering how to create an ECS cluster, specifying the VPC.

cluster, err := ecs.NewCluster(ctx, "test-cluster", &ecs.ClusterArgs{
			CapacityProviders: pulumi.StringArray{pulumi.String("FARGATE")},
			Name:              pulumi.String("test-cluster"),
			Settings:          nil,
			Tags:              nil,
		})

Where would the previously defined VPC be specified in the above?

Another question I'm having a tough time finding docs/examples on:

vpc, err := ec2.NewVpc(ctx, vpcName, vpcArgs)
        if err != nil {
            fmt.Println(err.Error())
            return err
        }

  ctx.Export("VPC-ID", vpc.ID())

In the above example,

vpc.ID()

is an

IDOutput

parameter. How do I convert it to an actual string so I can use the actual VPC ID in other Pulumi calls (i.e.

ec2.GetSubnetIds

)?

i think this documentation is wrong: https://www.pulumi.com/docs/reference/pkg/aws/apigateway/vpclink/ it says

<https://www.pulumi.com/docs/reference/pkg/aws/apigateway/vpclink/#target_arn_python|target_arn>

should be a list, but dosent work like that, i tested and it worked setting the arn as a str also https://www.pulumi.com/docs/reference/pkg/aws/lb/listener/ says that protocol "The protocol for connections from clients to the load balancer. Valid values are

TCP

,

TLS

,

UDP

,

TCP_UDP

,

HTTP

and

HTTPS

. Defaults to

HTTP

." but when testing i get:

* error creating ELBv2 Listener: ValidationError: Listener protocol 'HTTP' must be one of 'TLS, TCP, TCP_UDP, UDP'

Is there a full list of environment variables that are listened to? Mostly looking to setting region/zone/location for aws, gcp and azure out of env vars

Is there any guidance around managing multiple accounts / identities / organizations / mix of pulumi hosted and self-managed backends?

Is there a way to serialize a pulumi resource? I'm creating an api gateway, and for the api deployment to work correctly i need to define a trigger, which will be a sha of the serialized resources it depends on. Something similar to what is done on terraform: https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/api_gateway_deployment#redeployment-triggers or if someone has an example on another way to do it, i'll appreciate it.

Having a miserable time trying to debug a chart:

Error: invocation of kubernetes:yaml:decode returned an error: error converting YAML to JSON: yaml: mapping values are not allowed in this context
        at monitor.invoke (/tmp/issuer-phone-twilio/node_modules/@pulumi/pulumi/runtime/invoke.js:137:33)
        at Http2CallStream.call.on (/tmp/issuer-phone-twilio/node_modules/@grpc/grpc-js/src/client.ts:155:9)
        at Http2CallStream.emit (events.js:203:15)
        at process.nextTick (/tmp/issuer-phone-twilio/node_modules/@grpc/grpc-js/src/call-stream.ts:186:14)
        at process._tickCallback (internal/process/next_tick.js:61:11)

I'd share the whole thing but it has some secrets at the moment. What's even weirder is a previously working version of the chart is now giving the same error.

👋 I’m here! What’d I miss?

I changed my code to read the postgresql provider settings (host, admin username & pwd) no longer from Pulumi config but via a stackreference. After the changes, I get this as diff:

~ pulumi:providers:postgresql: (update)
        [id=6577980b-b7e1-466c-8a2d-18edaa0bc9fa]
        [urn=urn:pulumi:development::orthanc::pulumi:providers:postgresql::adminconnection]
        connectTimeout  : "180"
        database        : "postgres"
        databaseUsername: "psqladminun"
        host            : "<redacted>.<http://postgres.database.azure.com|postgres.database.azure.com>"
        password        : "[secret]"
        port            : "5432"
        superuser       : "false"
        username        : "psqladminun@<redacted>.<http://postgres.database.azure.com|postgres.database.azure.com>"

Pulumi reports a change to the provider, but none of the settings are different. Why reporting a change then?