Filed this bug on an issue we are having during our scheduled releases. Can you please have look?

Hello everyone. I have a question. Does anyone know how to save an object to DynamoDB? I am doing this:

const notificationTable = new aws.dynamodb.Table(
  "notificationTable",
  {
    readCapacity: 10,
    writeCapacity: 10,
    hashKey: "UUID",
    attributes: [
      {
        name: "UUID",
        type: "S",
      },
    ],
  }
);

const attributes = {
  name: "Name custom",
  item: [{ key: "key-custom", value: "value-custom" }],
};

const obj = {
  UUID: { S: "04b12ed0-b431-11eb-87d6-03d0311066e5" },
  CreatedAt: { S: "2021-05-13T21:20:37.275Z" },
  Event: { S: "custom_events" },
  NamedUser: { S: "<mailto:alex.quintero@gmail.com|alex.quintero@gmail.com>" },
  Attributes: { S: JSON.stringify(attributes) },
  Status: { S: "SEND" },
};

const notificationTableItem = new aws.dynamodb.TableItem(
  "notificationTableItem",
  {
    tableName: notificationTable.name,
    hashKey: notificationTable.hashKey,
    item: `${JSON.stringify(obj)}`,
  }
);

My result is:

Is it possible to make me stay as an Object? something like that:

{
  name: "Name custom",
  item: [{ key: "key-custom", value: "value-custom" }],
}

I was hoping to get some advice on how to use Pulumi to spin up Google Cloud Endpoints with OpenAPI auth (https://cloud.google.com/endpoints/docs/openapi/get-started-cloud-functions#reserve_hostname). The hard part is that it's a few steps and involves replacing a resource. Specifically, step 1 is to run something via Google's cloud run. It doesn't matter what container is run at this point; all that matters is we're "reserving" a unique hostname for our eventual cloud run service. Then we use the uniquely assigned hostname to create a text file that defines the routing and auth, then we include that generated text file in a docker image, and finally we replace the initial cloud run service with this new container. Since Pulumi's model is generally describe what you want and Pulumi will "make it so" this doesn't map neatly as we have to create a cloud run service that is temporary and we don't want it running when we're all done. The best idea I've come up with so far is a custom, dynamic resource provider that will create the temporary cloud run service and then "lie" in its

diff

method and such so that if it's been replaced it says not changes need to happen. I got that mostly working but it seems you can't use

pulumi_gcp

to create the temporary service inside the custom provider (I get an error that says, "Program run without the Pulumi engine available; re-run using the

pulumi

CLI" even though it was run via

pulumi up

). Even if I could use

pulumi_gcp

I don't see a way to destroy the resource created in the

create

method when

delete

is called. So I can make this work with native API calls or Google's own Python SDK but that all seems like more work than should be necessary. Is there a clean way to do this?

Hi friends, as you probably already know, the Stack Overflow annual developer survey is currently out. This is the first time that Pulumi has appeared in the survey. partypus We would love it if you ✅ the box for Pulumi if you haven't already filled out the survey. We would really appreciate the support! 🙏

https://github.com/pulumi/pulumi/issues/4032 — any reason why

--target-dependents

doesn’t actually work in simple situations, like the one demonstrated in the issue? I’m having and been having the same behavior since I started using Pulumi 2.10 and it exists in 3.3 still today. Is this just a feature that is half-working and poorly documented?

Anyone got a second to help me with this. Completely stuck. https://github.com/pulumi/pulumi/issues/7179

Hi Everyone, I am starting a new project where I basically need to deploy the “equivalent” of a bunch of REST (micro-) services, one or two databases an ingress-controller and kafka onto my k3s kubernetes cluster. Because I can choose to use the technology I want, I have decided to explore what it means to go “serverless”. Additionally, I want to use Pulumi as much as possible. I am a developer by trade, but had an active “devops”-role for over 5 years, with 24/7 on-call and all the shiny stuff. I basically want to accomplish what is shown in this tutorial https://www.pulumi.com/serverless/. Before I start, is this currently possible using Pulumi and a local k3s cluster?

Is Pulumi having a maintenance now? Deployments are AWFULLY slow for me 😕 Deploying just 4 lambdas and 8 queues is running now for over 2 hours without finishing (it also deletes 4 lambdas and 8 queues)

Hi everybody! Does somebody have experience with azure python function publishing with pulumi

azure_native

? Struggling to convert our ARM setup and cli replacement for

func azure functionapp publish --python [...]

🥲

Hi all, Can I get some advise on how to work around this https://github.com/pulumi/pulumi-awsx/issues/673? We have many routes in AWS API GW that could be configured via pulumi crosswalk

pulumi/awsx

lib's

awsx.apigateway.API

, but we also want the requestParams to be passed onto the downstream http target. Without that enhancement there should be some way to update the existing

IntegrationRoute

with these additional

requestParameters

Appreciate any help.

Can anyone explain what is the purpose of the

args: Inputs

argument of the

ComponentResource

constructor? From the docs it is not really clear.

new ComponentResource(type: string, name: string, args: Inputs, opts: ComponentResourceOptions, remote: boolean)

args

Information passed to [initialize] method.

When should we pass a value to

args

that is not the empty object?

Hey guys. Quick question on pricing: is there any limits for number of projects in the community license? edit: I wanna try a few layouts to deploy modules instead of the whole infrastructure as one single project.

Hey folks, Can anyone explain how to work with several libvirt hosts on the same stack ?

Anyone know if there’s a stack configuration that binds the awskms secret profile to a single/defined AWS_PROFILE, so that $AWS_DEFAULT_PROFILE doesn’t have to be set before invoking pulumi when using that secret provider?

This is important for situations where the aws profile used to encrypt/decrypt secrets may be different than the one used to run the aws providers in pulumi code.

I feel like the only alternative if you want to use awskms with a profile different than the main provider is to have all your pulumi code using the aws provider initialize its own resource provider and use a completely different profile configuration and pass that thing around to all ResourceOptions (yucky), simply so the awskms stack secrets provider can function with a separate profile.

Hi everyone After adding

eslint

to my project and format all files I run command

pulumi up

and receive multiple errors like this

aws:lambda:Function (dev-v1-internal-verify):

error: 1 error occurred:

* updating urn:pulumi:dev::aok-multi-pulumi::aws:lambda/function:Function::dev-v1-internal-verify: 1 error occurred:

* error modifying Lambda Function (dev-v1-internal-verify-cb6dab7) Code: RequestEntityTooLargeException:

status code: 413, request id: eadb73d4-c9b3-4946-8a4a-ed2743cde46b

Can anyone help ?

Hi all, just wanted to ask if following situation has a nice solution in Pulumi: I have a stateful resource A (shared block device) and stateless container resource B which works with the A (thus B depends on A). Besically the container B reads and writes stuff to the block device A. Now I want to reprovision A with a clean state. This generally means that I want to reprovision B as well to start with the newly provisoned A. But the problem is that when the A is reprovisioned, the old B is still up and basically corrupting the new A. I'd need the B to be deleted before the A is reprovisioned and only then create new B.

Hey all, I'm trying to get a secret into some userData (registering a gitlab runner and need to provide the registration token). Because I can't use the result of

config.requireSecret('gitlabRunnerRegistrationToken')

as a string I'm having a devil of a time. Originally I was going to use Mustache to render a template, but that's out. Using pulumi.concat is gonna be really ugly if I need to concat part of the script and another part of the script around the value...anyone have an idea of how to do this? Basically I'm taking a 26 line bash script and I need to get a Pulumi secret value into the middle of it then pass it to userData on my DO droplet Pulumi object.

in TypeScript

unless I could do the mustache rendering in an apply function? hmmm

I am running into a simple issue with a resource referring to a missing parent. How would I proceed in updating the rest of the stack?

Hi can anyone confirm this issue with setting new passphrase, it looks like when changing the passphrase , the command doesn't ask for the old passphrase. https://github.com/pulumi/pulumi/issues/6650#issuecomment-853618158

Hi, is there a way to execute something when a deployment was complete? We are using Pulumi type script.

Outside of the pulumi documentation, where do you guys look to find examples? or educational documentation? Currently using AWS/TS but anything helps.

Hi, I am running pulumi-python:2.25.2 on Jenkinspipeline to launch our stack but when I launch

pulumi login

command I got this error.

error: problem logging in: failed to get the home path: getting current user: luser: unable to get current user

. However, I have set the PATH variables like

export PATH=\$PATH:\$HOME/.pulumi/bin

any lead would be helpful here

how do you set up your Kubernetes provideR?

When using

pulumi login --local

, is there a way to skip/disable this paraphrase thing? There is absolutely no point in using it as there is no way it can ever be secure setting it as an env variable?

How can you pass overrides of the

values.yaml

in a Helm chart release? It only seems to take

Dictionary<string, object>

.