docs: document how to configure a custom endpoint resolver (#392)

Author: aajtodd

docs/howto/README.md

@@ -0,0 +1,8 @@
+# AWS SDK for Kotlin Guide
+
+This documentation supplements the official [AWS SDK for Kotlin Developer Guide](https://docs.aws.amazon.com/sdk-for-kotlin/latest/developer-guide/home.html).
+
+
+## Configuration
+
+* [Endpoints](./configuring/endpoints.md) - How to configure a custom endpoint resolver

docs/howto/configuring/endpoints.md

@@ -0,0 +1,27 @@
+# Configuring Client Endpoints
+
+The AWS SDK for Kotlin provides the ability to provide a custom endpoint to be used for a service. In most cases
+you can just use the default `EndpointResolver` provided with each service client. There are some reasons to provide 
+a custom endpoint though such as working with a pre-release version of a service or access to specific service 
+features not yet modeled in the SDK (e.g. S3 has dual-stack and FIPS endpoints).
+
+
+An [Endpoint Resolver](https://github.com/awslabs/aws-sdk-kotlin/blob/main/aws-runtime/aws-endpoint/common/src/aws/sdk/kotlin/runtime/endpoint/AwsEndpointResolver.kt#L11)
+can be configured to provide custom endpoint resolution logic for service clients. Every
+service client config is generated with an endpoint resolver that can be overridden. The endpoint resolver is given the
+service and region as a string, allowing for the resolver to dynamically drive its behavior. Each service client 
+package has an exported `ServiceId` constant that can be used to determine which service is invoking your endpoint 
+resolver.
+
+## Examples
+
+The following code snippet shows how a service endpoint resolver can be overridden for S3:
+
+```kotlin
+val sharedConfig = AwsClientConfig.fromEnvironment()
+val client = S3Client(sharedConfig) {
+    endpointResolver = AwsEndpointResolver { service, region ->
+       AwsEndpoint("https://mybucket.s3.us-west-2.amazonaws.com", CredentialScope(region = "us-west-2")) 
+    }
+}
+```