Add Retries section to README (#671)

## Summary

- Added a dedicated "Retries" section to the README, placed between
"Paginated responses" and "Single-Sign-On (SSO) with OAuth"
- The README introduction already mentions that the SDK performs
"intelligent retries," but there was no dedicated section explaining the
retry behavior
- Added corresponding TOC entry and updated NEXT_CHANGELOG.md

## Test plan

- [x] Verify the new TOC entry links correctly to the Retries section
- [x] Verify the Retries section renders correctly in GitHub markdown
- [x] Verify NEXT_CHANGELOG.md has the documentation entry

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

Author: renaudhartert-db

NEXT_CHANGELOG.md

@@ -12,6 +12,8 @@
 
 ### Documentation
 
+* Added "Retries" section to README.
+
 ### Internal Changes
 
 ### API Changes

README.md

@@ -13,6 +13,7 @@ The Databricks SDK for Java includes functionality to accelerate development wit
 - [Code examples](#code-examples)
 - [Long-running operations](#long-running-operations)
 - [Paginated responses](#paginated-responses)
+- [Retries](#retries)
 - [Single-sign-on with OAuth](#single-sign-on-sso-with-oauth)
 - [Error handling](#error-handling)
 - [Logging](#logging)
@@ -340,6 +341,10 @@ summary.stream()
    .forEach(jobSummary -> LOGGER.info("Latest: {}", jobSummary));
 ```
 
+## Retries
+
+The SDK automatically retries failed requests that are safe to retry. This includes HTTP 429 (Too Many Requests) and 503 (Service Unavailable) responses, as well as network-level errors such as connection resets and timeouts. The SDK uses a backoff strategy with jitter between retry attempts to avoid overwhelming the service. Retry behavior can be adjusted through the SDK configuration.
+
 ## Single-Sign-On (SSO) with OAuth
 
 ### Authorization Code flow with PKCE