diff --git a/core/sdk-core/src/main/java/software/amazon/awssdk/core/async/AsyncRequestBody.java b/core/sdk-core/src/main/java/software/amazon/awssdk/core/async/AsyncRequestBody.java
index 1bd7eecaeff2..0877557e35bd 100644
--- a/core/sdk-core/src/main/java/software/amazon/awssdk/core/async/AsyncRequestBody.java
+++ b/core/sdk-core/src/main/java/software/amazon/awssdk/core/async/AsyncRequestBody.java
@@ -375,6 +375,22 @@ static AsyncRequestBody fromRemainingByteBuffersUnsafe(ByteBuffer... byteBuffers
      * <p>An {@link ExecutorService} is required in order to perform the blocking data reads, to prevent blocking the
      * non-blocking event loop threads owned by the SDK.
      *
+     * <p><b>Executor Guidance:</b> The provided executor is used to run a blocking task that reads from the input stream and
+     * pushes data to the HTTP channel. If the executor has fewer threads than the number of concurrent requests using it,
+     * tasks are serialized — one slow or failing request may block other requests from writing data in a timely manner,
+     * leading to idle HTTP connections that the server may close before data can be written. This may result in an
+     * unrecoverable write timeout loop where every retry also fails.
+     *
+     * <p>To avoid this:
+     * <ul>
+     *     <li>Use a <b>dedicated</b> executor for SDK input stream requests — do not share it with other blocking work
+     *     in your application, as competing tasks may delay data writes and cause the same issue.</li>
+     *     <li>Size the executor's thread pool to at least the maximum number of concurrent requests.</li>
+     * </ul>
+     *
+     * <p>It is also recommended to configure an API call timeout via
+     * {@code ClientOverrideConfiguration.builder().apiCallTimeout()} to bound the total time spent on retries.
+     *
      * @param inputStream The input stream containing the data to be sent
      * @param contentLength The content length. If a content length smaller than the actual size of the object is set, the client
      *                      will truncate the stream to the specified content length and only send exactly the number of bytes
@@ -390,7 +406,9 @@ static AsyncRequestBody fromInputStream(InputStream inputStream, Long contentLen
 
     /**
      * Creates an {@link AsyncRequestBody} from an {@link InputStream} with the provided
-     * {@link AsyncRequestBodySplitConfiguration}.
+     * {@link AsyncRequestBodyFromInputStreamConfiguration}.
+     *
+     * <p>See {@link #fromInputStream(InputStream, Long, ExecutorService)} for guidance on executor.
      */
     static AsyncRequestBody fromInputStream(AsyncRequestBodyFromInputStreamConfiguration configuration) {
         Validate.notNull(configuration, "configuration");
diff --git a/core/sdk-core/src/main/java/software/amazon/awssdk/core/async/AsyncRequestBodyFromInputStreamConfiguration.java b/core/sdk-core/src/main/java/software/amazon/awssdk/core/async/AsyncRequestBodyFromInputStreamConfiguration.java
index c2d84af45632..ec8272458fb7 100644
--- a/core/sdk-core/src/main/java/software/amazon/awssdk/core/async/AsyncRequestBodyFromInputStreamConfiguration.java
+++ b/core/sdk-core/src/main/java/software/amazon/awssdk/core/async/AsyncRequestBodyFromInputStreamConfiguration.java
@@ -137,6 +137,10 @@ public interface Builder extends CopyableBuilder<AsyncRequestBodyFromInputStream
         /**
          * Configures the {@link ExecutorService} to perform the blocking data reads.
          *
+         * <p>It is recommended to have a dedicated executor for SDK input stream requests and have as many threads as the
+         * number of concurrent requests sharing it. Using an undersized or shared executor may lead to unrecoverable failures.
+         * See {@link AsyncRequestBody#fromInputStream(InputStream, Long, ExecutorService)} for details.
+         *
          * @param executor the executor
          * @return This object for method chaining.
          */