Hi, is there a way to let a

get*

function not run until another resource is created? I do not see the

depends_on

option for

InvokeOptions

. Thanks

There should be an

_output

variant of the get function. This allows arguments to be outputs. A workaround would be to use an output from a resource to defer the call, for example:

vpc = get_vpc_output(default=my_resource.id.apply(_: True))

Thanks for your reply. My usecase is:

import pulumi_gcp as gcp

project = gcp.organizations.Project(
    "project",
    name="project",
    project_id="project",
    billing_account="bills",
    auto_create_network=False,
)

vpc = gcp.compute.Network(
    "vpc",
    name="vpc",
    auto_create_subnetworks=False,
    project=project.name,
)

subnet = gcp.compute.Subnetwork(
    "subnet",
    name="subnet",
    network=vpc.id,
    region="us-central1",
    ip_cidr_range="10.0.0.0/8",
    private_ip_google_access=True,
    project=project.name,
)

gcp.compute.get_zones_output(
    region=subnet.region,
    project=subnet.project,
    status="UP",
)

but I still get the error that the project does not exist.

Does the error happen during preview?

yes

https://github.com/pulumi/pulumi/issues/9593 This seems related, basically pulumi is "assuming" the

name

output is the same as input, instead of treating the output use as a dependency, so uses it too early. I've seen it before, not really keen on this kind of optimisation. You should be able to do

project=subnet.id.apply(_: subnet.project)

Sure

Thanks, I've added some references to the issue