Could the SDK be updated to fail fast with a clear error message when it cannot access the IMDS?

[3pointer]

We are using the AWS SDK to obtain credentials for reading files from S3. The implementation works correctly in a local environment but fails in our cloud environment.

impl DefaultCredentialsProvider {
    async fn new(cli: impl HttpClient + 'static) -> Self {
        let cfg = ProviderConfig::default().with_http_client(cli);
        let default_provider = DefaultCredentialsChain::builder()
            .configure(cfg)
            .build()
            .await;
        Self { default_provider }
    }
}

We rely on a Node IAM role to grant access to S3. However, in the cloud environment, we encounter the following issue:

[6] [/workspace/.cargo/registry/src/mirrors.tuna.tsinghua.edu.cn-df7c3c540f42cdbd/aws-smithy-runtime-1.7.2/src/client/orchestrator.rs:172] [span_id=1099511627782] ["invoke"] [service="imds"] [operation="get-token"] [sdk_invocation_id=3874981] [elapsed=**12157.303551014s**]

This indicates that the call to IMDS (get-token) is not accessable after an excessively long duration (~3.4 hours).

We suspect that this issue is due to a misconfiguration, such as a missing sleep_impl. However, it would be better if the SDK reported an error instead of hanging indefinitely, as this behavior makes the issue harder to diagnose and resolve.

[3pointer]

After reviewing the code, I found that the issue is not related to the SDK. Instead, it stems from a bug caused by improper use of block_on with the Tokio runtime. Therefore, I will close the question for now.

[github-actions[bot]]

Hello! Reopening this discussion to make it searchable.