Bug: Race condition in WebClientStreamableHttpTransportErrorHandlingIT.testSessionRecoveryAfter404

Problem

The testSessionRecoveryAfter404 test has a race condition where an assertion runs before the previous async HTTP request completes, causing intermittent CI/CD failures.

Root Cause

Missing synchronization between async operations in the test's Mono chain. The test was asserting that a session ID was cleared before the HTTP request that invalidated the session had actually completed.

Solution

Add a 200ms delay between session loss handling and subsequent assertions to ensure proper async coordination (similar to existing pattern in test404WithSessionId).

Impact

This race condition has been causing intermittent CI/CD failures in the WebFlux transport integration tests.

Related PR

PR: feature/fix-mcp-transport-race-condition (galt-k fork)
Commit: be06d3d - "Fix race condition in MCP WebFlux transport error handling test"

Test Results

All 6 tests in WebClientStreamableHttpTransportErrorHandlingIT now pass consistently:

• test404WithoutSessionId ✓
• test404WithSessionId ✓
• test400WithoutSessionId ✓
• test400WithSessionId ✓
• testSessionRecoveryAfter404 ✓
• testReconnectErrorHandling ✓

Start with WebClientStreamableHttpTransportErrorHandlingIT.testSessionRecoveryAfter404 and compare its Mono chain with test404WithSessionId. Run the six named integration tests and confirm the recovery assertion consistently occurs after the async request completes.