hey folks, created* a PR, i hope it makes sense

i intend to write up a simple python\kubernetes example for the docs and also push these examples to examples repo?

sweet!

hey folks, any ideas how to use servicelist in kubernetes python package?

only reasonable reference I've found is: https://github.com/pulumi/pulumi-kubernetes/blob/master/tests/integration/get/step1/index.ts#L23

its like next level docs, what do I infer from this? 🙂

I have namespace defined at provider level and also doing this:

service_list = ServiceList(
        'list_k8s_services',
        items=[],
        metadata={
            "namespace": "kube-system"
        },    __opts__=ResourceOptions(provider=custom_provider)
    )

message has been deleted

hey, this is what is working for me for azure stuff:

i suppose you need to do roughly the same, define a async function and call your

get

inside of it with

await

hey folks, i'm trying to understand how to use depends on without passing in the resource variable

can I use resource urn or something? that I can construct easily without passing in resource variable throughout the configuration. the thing is, I've got "proxy" functions I'm calling, I could, use one more input parameter to pass to the proxy functions, but I'd like to avoid that, ideally

I think I foundsomething: https://github.com/pulumi/pulumi/blob/9eba0d3b378d5685e4312046cfb2a53542c5b70f/sdk/python/lib/pulumi/resource.py#L36 so I cant really pass anything except resource list?

Yes - in both Python and JavaScript the

dependsOn

resourceoption expects an array of resource objects. What's the scenario where you want/need to pass something else?

well, i know the urn i will get with the resource upfront, right? something like provider-type-name? so I dont really need to give it an object, because object I need to pass around

heh, I think I’m bumping into stuff as I explore 🙂 I’m trying to set up a peering connection between our management VPC (which is created and managed outside of pulumi/terraform) and a VPC that is part of a stack. All of the resource creation classes (eg ec2.Vpc, ec2.Instance etc) take an options that include the possibility of an alternate provider. But the query functions (eg get_vpc) do not.

Basically, all of the query function, which are the equivalent of data sources in terraform, need to take opts in the same way as the resource classes.

I’ve created an issue in github, https://github.com/pulumi/pulumi-aws/issues/466 not sure if that’s the right place 😄

hey folks, any feedback on my PR's?

Hey, it’s a holiday today and not many people are working - we’ll get on it tomorrow

oh, ok. sorry 😉

hey folks, i've fixed most of the comments except for one that doesnt make a lot of sense to me ^) there's one more pending pr 🙂

I hit an interesting issue today. I was setting up routing tables and subnets and needed to associate each subnet to its appropriate routing table. The issue was having to create a unique name for each association, even though there was no reason to assign the association to a variable for later use. So what I wanted was something like a resource name of “{}-{}“.format(route_table.id, subnet.id) but they are, of course, Outputs. The resource name has to be a string when the object is initialized, so can’t use Outputs. But the inidividual resource names for the route table and the subnet are already in place and aren’t (or maybe shouldn’t be) Outputs. Is there a way to expose runtime attributes of an object that aren’t dependent on resolution during processing, like the resource name?

i think you need to use apply

redis_connstr = redis.apply(
        lambda redis: "{}.<http://redis.cache.windows.net:6380,password={},ssl=True,abortConnect=False%22.format(|redis.cache.windows.net:6380,password={},ssl=True,abortConnect=False".format(>
            BRANCH, redis)
    )

where redis is a output of a resource

but, you decide the resource name, i'm just using a prefix + suffix (equals to resource type) everywhere