If I create a password with

random.RandomPassword

, and then use it for something like a database admin password. Is there a way that I can get that password if I want to login to the DB manually? Similar to a password manager like Bitwarden/Lastpass. The docs only explains how it is stored in state but I can't see anything on how you can view that password.

https://github.com/pulumi/pulumi-awsx/pull/643 any reviewers on this?

Is there a design pattern for replacing a resource when a config var is

true

? I'd like to have a 'password reset' configuration option for some secrets in AWS secrets manager that would, when set, trigger a rotation. Was thinking about generating a random string to append to the resource name, or creating a dynamic resource that the rotation resource is dependent on, but both approaches have a couple of problems

Is there a way to create a fake StackReference for unit tests?

Is there a way in unit testing to await all resources?

Is it possible to tag an aws resource after its creation? I need to create a subnet, pass it to a pulumi_eks.Cluster(), then tag the subnet with the cluster name. thanks.

Is there an example for writing custom components as good as this, but for Go? https://github.com/pulumi/examples/blob/master/aws-ts-serverless-datawarehouse/datawarehouse/index.ts I've only found simple components, but curious to see how people structure components scope.

If I create a new eks cluster in a stack, how do I switch to that cluster to then apply a k8s resource? Is it enough to simply create a Provider(cluster=cluster) resource? k8s.yaml.ConfigFile doesn't seem to take a provider argument. Thanks.

I’ve been combing through the typescript source code and the documentation trying to find an example of unit testing a Component Resource, but the examples I stumble across seem to be for direct resource uses in a file, and not say exporting the Component Resource class and initialising it. Does anyone happen to know where I can find an example/documentation for this?

is the only way to create an azure ad application using the AzureAD provider (that is, you can’t do it with azure-nextgen?)

This message was deleted.

I'm trying to apply a bunch of k8s resources from a single file with k8s.yaml.ConfigFile(). Is there any guarantee that they are applied in order as read from the file? I'm noticing e.g. CRDs not being applied first which is causing other things to fail. Thanks.

Is there a way of knowing if Pulumi is going to destroy a resource prior to running

pulumi up

? When I run preview it simply says

update

, with no indication of what impact the update will have. My use case is that I am running on PaaS (App Service & servless in Azure) and I want to be able to run updates against the live infrastructure. In order to do this safely, I need to assess if resources will be destroyed & recreated or merely updated.

Perhaps I need to rephrase my earlier question: What are some good technique(s) for either a) ensuring resources aren't destoryed or b) warning/stopping before destroying resources?

Hey, I have an issue with cloudwatch metrics creation. I have

rds.ts

where I create an AWS RDS instance :

import * as config from "./config"
import * as iam from "./iam"
import * as secrets from "./secrets"
import * as vpc from "./vpc"

const dbSubnets = new aws.rds.SubnetGroup("dbsubnets", {
    subnetIds: vpc.vpc.privateSubnetIds,
});

let analyticsDBConf = config.services.aws.analyticsdb
export const instance = new aws.rds.Instance("analyticsdb", {
    engine: "postgres",
    engineVersion: "12.5",
    
    instanceClass: analyticsDBConf.instanceType!,
    
    name: "DBName",
    username: "DBMasterUser",
    password: secrets.analyticsdb.passsword.result,
    
    dbSubnetGroupName: dbSubnets.name,
    vpcSecurityGroupIds: [
        vpc.sgAnalyticsDB!.id
    ],
    domainIamRoleName: iam.servicesInstanceRole.name,
    
    allocatedStorage: 20,
    maxAllocatedStorage: 100,
    backupRetentionPeriod: 7,
    backupWindow: "05:18-05:48",
    skipFinalSnapshot: true,
});

Then, in

cloudwatch.ts

, I try to create cloudwatch metric alarm on the RDS instance's free disk space :

import * as analyticsdb from "./rds"
import * as sns from "./sns"
import * as config from "./config"
import * as aws from "@pulumi/aws"

new aws.cloudwatch.MetricAlarm("analyticsdb-storage", {
    alarmActions: [sns.alarmsTopic],
    alarmDescription: "Monitoring space usage on analytics DB",
    comparisonOperator: "GreaterThanThreshold",
    datapointsToAlarm: 10,
    evaluationPeriods: 10,
    metricName: "FreeStorageSpace",
    namespace: "AWS/RDS",
    dimensions: {
        DBInstanceIdentifier: analyticsdb.instance.identifier
    },
    okActions: [sns.alarmsTopic],
    period: 60,
    statistic: "Maximum",
    threshold: 90,
    tags: {
        deployment: config.deploymentName,
    },
}, {
    dependsOn: [
        analyticsdb.instance
    ],
})

But when I do

pulumi up

I keep getting the following exception on preview, even though I've added an explicit dependency on the RDS instance :

TypeError: Cannot read property 'identifier' of undefined
        at Object.<anonymous> (/home/ort/work/private-cloud/cloudwatch.ts:15:52)

Did Pulumi change its search functionality on the docs site?

Does anyone know where the repo for https://github.com/joeduffy/project-cloudysky is? Related to

@flat-insurance-25294 it looks from the video to be a private repo

Ah.

Well I had a question regarding that video.

How would you separate the stacks between

infra

and

app

When dealing with CI while keeping one project on pulumi and one github repo?

if you are looking at using our GitHub Action, please be aware - in about 1 hour - we are publishing a new major version that has changed to be native to Actions rather than docker based

ah I preferred docker based ones!

the docker based action will be deprecated and be in use for 6 more months

Is there a way to run pulumi up without a stack and create it in code? LIke ephemeral stacks, the docs mentioned it but never said how

@broad-dog-22463 Does that mean that I can’t reference pulumi workflow in my own docker image? The reason is, we got our on image for running on Github actions, it runs some scripts that eventually triggers Pulumi

pulumi supports aws timestream ?

What if I don’t use Pulumi actions, but make my own and just install pulumi CLI?

our Pulumi docker images will not be going away - just the action based on docker