[BUG] Amazon S3 Bucket connector not working

[karzuo]

Type of Connector

Independent Publisher Connector

Name of Connector

Amazon S3 Bucket (Independent Publisher)

Describe the bug

It always throw an internal server error whichever method I tried:

{
    "statusCode": 500,
    "headers": {
        "Access-Control-Allow-Methods": "GET, PUT, PATCH, DELETE, POST",
        "Access-Control-Allow-Origin": "*",
        "Access-Control-Max-Age": "3600",
        "Access-Control-Expose-Headers": "*",
        "x-ms-environment-id": "default-2414a4c8-fba7-469b-acad-5dfec2865467",
        "x-ms-tenant-id": "2414a4c8-fba7-469b-acad-5dfec2865467",
        "x-ms-dlp-re": "-|-",
        "x-ms-dlp-gu": "-|-",
        "x-ms-dlp-ef": "-|-/-|-|-",
        "Date": "Wed, 26 Mar 2025 01:34:31 GMT",
        "Content-Length": "419",
        "Content-Type": "application/json"
    },
    "body": {
        "error": {
            "code": 500,
            "message": "Sequence contains no elements\r\n   at System.Linq.Enumerable.First[TSource](IEnumerable`1 source)",
            "source": "default-2414a4c8-fba7-469b-acad-5dfec2865467.03.common.asia.azure-apihub.net",
            "path": "choose[4]\\when[1]\\choose[1]\\otherwise",
            "policyId": "",
            "clientRequestId": "92735caa-7d10-4520-be4f-4ee59dd76ea3",
            "dreeSet": "true"
        }
    }
}

Is this a security bug?

No, this is not a security bug

What is the severity of this bug?

Severity 1 - Connector is broken or there's a very serious issue

To Reproduce

Set up a simple manual trigger pipeline with Delete Object.

Region: ap-southeast-1

Expected behavior

It should return status 200, success regardless whether the file is present (I tried it by using http request, manually signing the api and it worked)

Environment summary

Set up via Cloud Flow UI:

Additional context

Not sure if it is a connection issue, because the error is the same when I try giving wrong credentials as well. There is no issue with my credentials as it has full access to S3 and tested to work fine using other methods.

[vmanoharas]

Tagging owner of the Amazon S3 Bucket connector @megel

[RLSid17]

It also affects my flow using this connector it started last March 18 despite having good connections and credentials. This is indeed a bug. I do hope this will be fixed the soonest.

[megel]

@RLSid17 I have checked in my VAL environment with the current actions:

Can you please provide detailed information about the parameters that you have used with the action. Especially the path (how it is structured)

[RLSid17]

@RLSid17 I have checked in my VAL environment with the current actions:

Can you please provide detailed information about the parameters that you have used with the action. Especially the path (how it is structured)

This is the parameters set for it:

There is no folder path, only the bucket and the object. for eg. bucket/key.csv

Other than the Delete Object, I tried List Objects and it didn't work:

The Delete Object worked before March 18 so I'm really not sure what the cause is. As what @karzuo reported, we have the same issue and error.

[megel]

@RLSid17 this is really crazy. Which region is your power platform environment?

I have run all 4 actions:

I also used a simplified filename:

I'm unfortunately a bit limited in my testing environments. 😥 I have no sponsor. Therefore, I use usually a developer environment.
However, my session and cluster information are these:

Cluster environment: Prod
Cluster category: Prod
Cluster geo name: DE
Cluster URI suffix: de-il101.gateway.prod.island

I assume you are in US right? Maybe someone from MSFT @vmanoharas can help to debug and locate the issue 😉

[RLSid17]

@RLSid17 this is really crazy. Which region is your power platform environment?

I have run all 4 actions:

I also used a simplified filename:

I'm unfortunately a bit limited in my testing environments. 😥 I have no sponsor. Therefore, I use usually a developer environment. However, my session and cluster information are these:

Cluster environment: Prod
Cluster category: Prod
Cluster geo name: DE
Cluster URI suffix: de-il101.gateway.prod.island

I assume you are in US right? Maybe someone from MSFT @vmanoharas can help to debug and locate the issue 😉

The region of the environment is Asia. Please help with this issue. It is already affecting our data with duplicate records and we have to manually delete the objects which is really not efficient in our case.

[megel]

@karzuo @RLSid17 I have setup my test flow with nearly the same parameters like yours:

I also used a Bucket that is located in ap-southeast-1 and this worked.

The only difference might be that my Dataverse Environment is in Germany / Europe. Where is your Dataverse environment. Please provide your session details including the cluster and region.

My session details are this:

[RLSid17]

@karzuo @RLSid17 I have setup my test flow with nearly the same parameters like yours:

I also used a Bucket that is located in ap-southeast-1 and this worked.

The only difference might be that my Dataverse Environment is in Germany / Europe. Where is your Dataverse environment. Please provide your session details including the cluster and region.

My session details are this:

This is the session details:

Timestamp: 2025-04-01T00:56:29.348Z
Build name: 0.0.20250321.5-2503.3-prod
Unique name: org7137d82a
Environment ID: Default-2576a09d-a3f6-4a63-a170-c7ff95ea93ce
Cluster environment: Prod
Cluster category: Prod
Cluster geo name: AS
Cluster URI suffix: as-il101.gateway.prod.island

Did another test, still the same:

@karzuo Are you still having the same issue?

[megel]

@RLSid17 I guess, this is an issue at the region Asia.

I can confirm, that I can repro the issue now at Asia. After creating a fresh Developer Environment in region Asia the connector run into similar issues:

@vmanoharas please confirm, that the connector is rolled out correctly the region Asia?

Here are my session details:

Timestamp: 2025-04-01T07:14:49.017Z
Session ID: 0fc77a10-0ec6-11f0-bc27-a7f337b28cdb
Tenant ID: 0561be39-5ca4-4694-9b53-8f7ed78b9d61
Object ID: bc333f13-b56a-462b-89ec-c572df5cabc3
Build name: 0.0.20250329.1-2503.4-prod
Organization ID: 67f62ca4-420e-f011-9aed-000d3a80b290
Unique name: unq67f62ca4420ef0119aed000d3a80b
Instance url: https://org58cdf356.crm5.dynamics.com/
Environment ID: 77b6e8df-e492-eb2e-af5d-46fa6bef1f9a
Cluster environment: Prod
Cluster category: Prod
Cluster geo name: AS
Cluster URI suffix: as-il101.gateway.prod.island

[vmanoharas]

Yes @megel Amazon S3 Bucket deployed to all the regions without any issues. Please refer the screen shot below.

[RLSid17]

Issue still the same:

Surely there must be a solution to this?

[megel]

Yes @megel Amazon S3 Bucket deployed to all the regions without any issues. Please refer the screen shot below.

@vmanoharas yes, I see that the connector is deployed to Asia. But what is different is the behavior to all other environments. I assume, that something might be different with the custom code which the connector is using.

The problem is, I do not have insights of what happens in the backend systems.

There must be something different in this region!
I have importing the custom connector from the GitHub code into this region and added it to my test flow.

The result is, that the imported connector works as expected 😮:

Independent Publisher version is left ... still failing
Imported connector version is right ... works as expected.

@vmanoharas how can we dive deeper into this issue? I guess here can only MSFT help.

The issue is obviously only in this region present.

[vmanoharas]

Yes @megel Amazon S3 Bucket deployed to all the regions without any issues. Please refer the screen shot below.

@vmanoharas yes, I see that the connector is deployed to Asia. But what is different is the behavior to all other environments. I assume, that something might be different with the custom code which the connector is using.

The problem is, I do not have insights of what happens in the backend systems.

There must be something different in this region! I have importing the custom connector from the GitHub code into this region and added it to my test flow.

The result is, that the imported connector works as expected 😮:

Independent Publisher version is left ... still failing Imported connector version is right ... works as expected.

@vmanoharas how can we dive deeper into this issue? I guess here can only MSFT help.

The issue is obviously only in this region present.

Hi @megel,

Thank you for your response. I would recommend to create a support ticket with MSFT, respective deployment team would address this. @RLSid17, kindly create a support ticket with this issue.

[RLSid17]

Yes @megel Amazon S3 Bucket deployed to all the regions without any issues. Please refer the screen shot below.

@vmanoharas yes, I see that the connector is deployed to Asia. But what is different is the behavior to all other environments. I assume, that something might be different with the custom code which the connector is using.
The problem is, I do not have insights of what happens in the backend systems.
There must be something different in this region! I have importing the custom connector from the GitHub code into this region and added it to my test flow.
The result is, that the imported connector works as expected 😮:

Independent Publisher version is left ... still failing Imported connector version is right ... works as expected.
@vmanoharas how can we dive deeper into this issue? I guess here can only MSFT help.
The issue is obviously only in this region present.

Hi @megel,

Thank you for your response. I would recommend to create a support ticket with MSFT, respective deployment team would address this. @RLSid17, kindly create a support ticket with this issue.

Can you please redirect me where do I raise this? Or maybe you can raise on behalf of those who are affected?