Pulumi keep updating chart with no change

[bglgwyng]

What happened?

I deployed a helm chart for aws alb controller on my eks cluster. Pulumi succeeded at creating resources and deploying the chart. Then I ran pulumi up again without any change in my pulumi code. However, Pulumi detected changes in resources.

 +-  │     ├─ kubernetes:core/v1:Secret                                                  default/aws-load-balancer-tls  replace    
 ~   │     ├─ kubernetes:admissionregistration.k8s.io/v1:MutatingWebhookConfiguration    aws-load-balancer-webhook      update     
 ~   │     └─ kubernetes:admissionregistration.k8s.io/v1:ValidatingWebhookConfiguration  aws-load-balancer-webhook      update 

And this (maybe) redundant update also got succeded. But every time I ran pulumi up, it kept reporting the same changes to me and updating it.

Steps to reproduce

I made a small custom resource AlbIngressController.ts. You can deploy to your aws eks cluster with this code

new AlbIngressController('alb-ingress-controller', { cluster });

In my case, cluster was a fargate cluster. And run pulumi up and see it succeed. The run pulumi up and see the changes Pulumi reports.

Expected Behavior

At the second time I ran pulumi up without any change in my code, Pulumi should say it had nothing to update.

Actual Behavior

Pulumi keeps reporting the same change and updating the resources.

Versions used

CLI          
Version      3.37.2
Go Version   go1.18.4
Go Compiler  gc

Plugins
NAME        VERSION
aws         5.10.0
cloudflare  4.9.0
docker      3.2.0
eks         0.41.2
kubernetes  3.20.3
nodejs      unknown
random      4.7.0

Host     
OS       darwin
Version  11.6
Arch     x86_64

This project is written in nodejs: executable='/usr/local/bin/node' version='v16.13.2'

Additional context

The first time I deployed aws alb controller chart to the cluster, Pulumi said the deployment was successful but the pods in the cluster were not healthy state. It reported warning that said 'Back-off restarting failed container'.

Contributing

Vote on this issue by adding a 👍 reaction. To contribute a fix for this issue, leave a comment (and link to your pull request, if you've opened one already).

[gitfool]

One way to avoid this churn is to install cert manager and enable it for aws load balancer controller.

That said, I would interested to know if there are other ways to avoid this churn without cert manager.

[guineveresaenger]

Hi @bglgwyng - can you please send us a complete program of what you saw this behavior with, so we can reproduce this behavior? Thank you!

[bglgwyng]

@guineveresaenger

https://github.com/bglgwyng/pulumi-k8s-report

I made a smaller one that can reproduce the same behavior.

[k1ng440]

use k8s.helm.v3.Release instead of k8s.helm.v3.Chart