Pulumi state does not reflect actual resource state

[sfc-gh-enosenko]

What happened?

When pulumi ran apply it failed to apply resource spec update but still went ahead and updated it's state. The instance was having some performance issues yet after several bump-ups in specs we noticed that the changes were not applied.

Pulumi state shows values that were defined in code:

Disk size - 400GB
IOPS - 12000

Yet the instance is still running with old configuration

Pulumi SDK version - v3.137.0

Example

pulumi state for the resource

{
    "__meta": "{\"e2bfb730-ecaa-11e6-8f88-34363bc7c4c0\":{\"create\":2400000000000,\"delete\":3600000000000,\"update\":4800000000000},\"schema_version\":\"2\"}",
    "allocatedStorage": 400,
    // ...
    "iops": 12000,
    "password": {

    },
    // ...
}

Output of pulumi about

pulumi about
CLI          
Version      3.146.0
Go Version   go1.23.4
Go Compiler  gc

Plugins
KIND      NAME        VERSION
resource  aws         6.59.1
language  go          3.146.0
resource  kubernetes  3.30.2
resource  random      4.13.2

Host     
OS       darwin
Version  15.3
Arch     arm64

This project is written in go: executable='/opt/homebrew/opt/go/libexec/bin/go' version='go version go1.23.5 darwin/arm64'

Backend        
Name           pulumi.com
URL            https://app.pulumi.com/XXX
User           XXX
Organizations  XXX, XXX
Token type     personal

Dependencies:
NAME                                        VERSION
github.com/apparentlymart/go-cidr           v1.1.0
github.com/aws/aws-sdk-go-v2                v1.24.0
github.com/aws/aws-sdk-go-v2/config         v1.25.6
github.com/aws/aws-sdk-go-v2/credentials    v1.16.5
github.com/aws/aws-sdk-go-v2/service/ec2    v1.137.2
github.com/aws/aws-sdk-go-v2/service/eks    v1.35.5
github.com/aws/aws-sdk-go-v2/service/iam    v1.27.4
github.com/aws/aws-sdk-go-v2/service/sts    v1.25.5
github.com/aws/smithy-go                    v1.19.0
github.com/go-playground/validator/v10      v10.20.0
github.com/hashicorp/go-version             v1.7.0
github.com/pkg/errors                       v0.9.1
github.com/pulumi/pulumi-aws/sdk/v6         v6.59.1
github.com/pulumi/pulumi-kubernetes/sdk/v3  v3.30.2
github.com/pulumi/pulumi-random/sdk/v4      v4.13.2
github.com/pulumi/pulumi/sdk/v3             v3.137.0
github.com/sirupsen/logrus                  v1.9.3
github.com/snowflakedb/k8s-go-kit           v0.2.1-0.20250207221356-7705722259e2
github.com/stretchr/testify                 v1.9.0
gopkg.in/yaml.v3                            v3.0.1
k8s.io/apimachinery                         v0.27.1

Pulumi locates its logs in /var/folders/w3/m48z_ys17p9fvr_hh78ympb80000gn/T/ by default
warning: Failed to get information about the current stack: No current stack

in go.mod:

github.com/pulumi/pulumi/sdk/v3 v3.137.0

Additional context

No response

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).

[Frassle]

This is probably an issue with the aws provider rather than the engine, so going to move repos.

[sfc-gh-enosenko]

@Frassle a bit more context

instance is running gp3

the chronology of changes:

• iops bumped up to 6000 ( i wonder why pulumi didn't fail, because for the storage size ( 100gb ), max possible iops is capped at 3000 )
• disk size bumped up to 400gb ( again, pulumi should've failed because min iops starts at 12000 and it was set by previous commit to 6000 )
• iops bumped up to 12000