pulumi-community #python

python

sparse-state-34229

06/11/2020, 7:08 PM

there’s a lot more in https://github.com/pulumi/examples

chilly-hairdresser-56259

06/11/2020, 7:15 PM

@sparse-state-34229 I have been looking at many examples and documentation and haven't figured out how to properly manipulate the strings. For one thing it works for another item such as this if I copy and paste with a few minor changes it fails.

sparse-state-34229

06/11/2020, 7:15 PM

yeah, it’s kind of annoying like that 😞

chilly-hairdresser-56259

06/11/2020, 7:16 PM

So if Output and Apply do not work does that mean its simply not possible to reference it?

sparse-state-34229

06/11/2020, 7:17 PM

they do work

👍 1

sparse-state-34229

06/11/2020, 7:17 PM

it’s likely just a matter of using them the right way for this use case

chilly-hairdresser-56259

06/11/2020, 7:17 PM

gotcha.

sparse-state-34229

06/11/2020, 7:19 PM

how are you trying to use

service[0].name

sparse-state-34229

06/11/2020, 7:19 PM

eg as a function argument?

chilly-hairdresser-56259

06/11/2020, 7:21 PM

service[0].name

would be the output from the ecs service. I need to get string format of the output for

resource_id

in the appautoscaling. Format

service/ecs cluster name/ecs service name

. Yes Function argument

sparse-state-34229

06/11/2020, 7:27 PM

so eg

do_something(service_name=service[0].name)

yeah?

sparse-state-34229

06/11/2020, 7:28 PM

if so, try calling

do_something()

within the apply lambda

chilly-hairdresser-56259

06/11/2020, 7:29 PM

appautoscalingTarget = aws.appautoscaling.Target(
        resource_name= "%s-%s-%s-asg-target" % (p_env, tasks[x]['application'], tasks[x]['service']),
        min_capacity= tasks[x]['minimumCapacity'],
        max_capacity= tasks[x]['maximumCapacity'],
        resource_id= f"service/{service[0].name}/{ecs_cluster.name}",
        scalable_dimension= "ecs:service:DesiredCount",
        service_namespace= "ecs"
    )

chilly-hairdresser-56259

06/11/2020, 7:30 PM

being the iterator of my list[dict{}] of ecs tasks configs

sparse-state-34229

06/11/2020, 7:32 PM

what do you do with

appautoscalingTarget

chilly-hairdresser-56259

06/11/2020, 7:35 PM

appautoscalingTarget

I use create the scaling policies. So I then reference like such in the scaling policies:

appautoscalingTarget.scalable_dimension

for the resource

aws.appautoscaling.Policy

sparse-state-34229

06/11/2020, 7:36 PM

okay

sparse-state-34229

06/11/2020, 7:37 PM

you might need to move the output massaging up to

aws.appautoscaling.Policy

gentle-diamond-70147

06/11/2020, 7:40 PM

What is

service

here?

chilly-hairdresser-56259

06/11/2020, 7:40 PM

service

= ecs Service.

service

creates multiple resources and returns two objects. ECS Service and taskDefinition

gentle-diamond-70147

06/11/2020, 7:47 PM

I believe this should work for you:

appautoscalingTarget = aws.appautoscaling.Target(
        resource_name= "my-target",
        min_capacity= 1,
        max_capacity= 2,
        resource_id= Output.all(service[0].name, ecs_cluster.name).apply(lambda args: f"service/{args[0]}/{args[1]}"),
        scalable_dimension= "ecs:service:DesiredCount",
        service_namespace= "ecs"
    )

gentle-diamond-70147

06/11/2020, 7:48 PM

I changed the other inputs to test in my code, so only copy the

resource_id

input line :)

gentle-diamond-70147

06/11/2020, 7:52 PM

As Scott said, you can't print the output of apply, but you can print within the apply. e.g.

Output.all(service[0].name, ecs_cluster.name).apply(lambda args: print(f"service/{args[0]}/{args[1]}"))

gentle-diamond-70147

06/11/2020, 7:52 PM

.apply()

always returns an Output

chilly-hairdresser-56259

06/11/2020, 7:54 PM

Ok, so somewhere I must have messed up when prevously trying to get this to work with

Output.all().apply()

I am going to try to figure out why.

chilly-hairdresser-56259

06/11/2020, 7:55 PM

When referencing

output.all().apply()

does that implicitly "wait" for the previous resource output, making it like a

depends_on

chilly-hairdresser-56259

06/11/2020, 7:55 PM

gentle-diamond-70147

06/11/2020, 7:57 PM

Effectively, yes. Pulumi will wait for the outputs passed into

.all()

to be come "real" before proceeding with the lambda.

👍 1

gentle-diamond-70147

06/11/2020, 8:02 PM

Oh... I think you need to reverse the arguments to

.all()

. It looks like AWS expects

service/<cluster-name>/<service-name>

gentle-diamond-70147

06/11/2020, 8:02 PM

Output.all(ecs_cluster.name, service[0].name).apply(lambda args: f"service/{args[0]}/{args[1]}")