Cae 633 add functional tests notifications

[kiryazovi-redis]

Make sure that:

• You have read the contribution guidelines.
• You have created a feature request first to discuss your contribution intent. Please reference the feature request ticket number in the pull request.
• You applied code formatting rules using the mvn formatter:format target. Don’t submit any formatting related changes.
• You submit test cases (unit or integration tests) that back your changes.

---

CAE-633: Add Functional Tests for Redis Enterprise Push Notifications

Overview

This PR adds comprehensive functional tests to validate Redis Enterprise push notifications during maintenance events, ensuring the Lettuce client correctly captures and processes RESP3 push messages for cluster state changes.

What's Added

• Complete test suite for all Redis Enterprise push notification types:

  • MOVING - Endpoint migration notifications
  • MIGRATING - Data migration in progress
  • MIGRATED - Data migration completed
  • FAILING_OVER - Failover initiation
  • FAILED_OVER - Failover completion

• Dynamic cluster discovery with real-time adaptation to cluster state changes

• Proper RESP3 push message capture using Lettuce's PushListener API

• Robust test isolation with state recording and restoration between tests

Technical Implementation

Push Notification Capture:

• Replaced flawed ping-based approach with proper PushListener API integration
• Added ByteBuffer decoding to extract actual IP addresses from push messages
• Successfully hooks into Lettuce's protocol layer to capture RESP3 push messages

Cluster State Management:

• Implemented RedisEnterpriseConfigDiscovery with cache clearing for fresh master/slave lists
• Added dynamic node selection for failover operations
• Real-time cluster configuration discovery prevents stale state issues

Fault Injection:

• Enhanced FaultInjectionClient with proper validation and retry logic
• Added support for migration operations with appropriate timeouts (300s)
• Fixed infinite retry loops in validation

Critical Bug Fixes Applied

1. Double-Prefixing Bug: Fixed malformed shard IDs (redis:redis:1 → redis:1) in state recording
2. Migration Timeouts: Increased timeout from 120s to 300s for data migration operations
3. Restoration Logic: Fixed cluster state restoration to only failover master shards (Redis Enterprise constraint)

Test Results

Tests run: 5, Failures: 0, Errors: 0, Skipped: 0
Total time: 09:05 min

All notification types successfully validated with proper push message capture and cluster state management.

Files Changed

• MaintenanceNotificationTest.java - Main test suite with 5 comprehensive tests
• FaultInjectionClient.java - Enhanced fault injection with proper timeouts
• RedisEnterpriseConfigDiscovery.java - Dynamic cluster discovery implementation
• RedisEnterpriseConfig.java - Cluster configuration models
• Additional utility classes for test infrastructure

Benefits

• Production readiness: Validates real-world Redis Enterprise maintenance scenarios
• Proactive monitoring: Ensures applications can react to cluster state changes
• Regression prevention: Comprehensive test coverage prevents future push notification issues
• Dynamic adaptation: Tests work with any Redis Enterprise cluster configuration

Testing

Tests validated against live Redis Enterprise clusters with real maintenance operations. The test suite adapts dynamically to different cluster configurations and provides comprehensive coverage of all push notification scenarios.

---

This PR establishes a solid foundation for Redis Enterprise push notification testing, ensuring the Lettuce client maintains robust support for maintenance event handling in production environments.

[tishun]

Good, but I think we can refactor the code a bit to make it more readable and easier to maintain

[Copilot]

Pull Request Overview

This PR adds comprehensive functional tests for Redis Enterprise push notifications during maintenance events, enabling the Lettuce client to properly capture and validate RESP3 push messages for cluster state changes. The implementation includes full test coverage for all notification types (MOVING, MIGRATING, MIGRATED, FAILING_OVER, FAILED_OVER) with dynamic cluster discovery and proper state management.

• Complete test suite validating all Redis Enterprise push notification types during maintenance operations
• Dynamic cluster configuration discovery with real-time adaptation to cluster state changes
• Enhanced fault injection client with proper timeout management and validation for Redis Enterprise operations

Reviewed Changes

Copilot reviewed 4 out of 5 changed files in this pull request and generated 5 comments.

File	Description
MaintenanceNotificationTest.java	Main test suite with 5 comprehensive tests for all push notification types
RedisEnterpriseConfig.java	Dynamic cluster discovery and configuration management with validation
FaultInjectionClient.java	Enhanced fault injection with proper timeouts, validation, and Redis Enterprise operation support
FaultInjectionClientUnitTest.java	Unit tests for fault injection client validation and basic functionality

Comments suppressed due to low confidence (2)

[Copilot]

This method and extractNumericNodeId() have similar logic but handle null differently. Consider creating a generic extract method or ensuring consistent null handling patterns.

Suggested change

	* Extract the numeric part of shard ID from full format "redis:X" -> "X"
	*/
	private String extractNumericShardId(String fullShardId) {
	if (fullShardId == null) {
	return null;
	}
	if (fullShardId.contains(":")) {
	return fullShardId.split(":")[1];
	}
	return fullShardId;
	}
	/**
	* Extract the endpoint ID part from full format "endpoint:X:Y" -> "X:Y"
	*/
	private String extractEndpointId(String fullEndpointId) {
	if (fullEndpointId == null) {
	return null;
	}
	if (fullEndpointId.startsWith("endpoint:")) {
	return fullEndpointId.substring("endpoint:".length());
	}
	return fullEndpointId;
	* Extract a specific part of a string based on a prefix and optional delimiter.
	*/
	private String extractPart(String input, String prefix, String delimiter) {
	if (input == null) {
	return null;
	}
	if (prefix != null && input.startsWith(prefix)) {
	input = input.substring(prefix.length());
	}
	if (delimiter != null && input.contains(delimiter)) {
	return input.split(delimiter)[1];
	}
	return input;
	}
	/**
	* Extract the numeric part of shard ID from full format "redis:X" -> "X"
	*/
	private String extractNumericShardId(String fullShardId) {
	return extractPart(fullShardId, "redis:", ":");
	}
	/**
	* Extract the endpoint ID part from full format "endpoint:X:Y" -> "X:Y"
	*/
	private String extractEndpointId(String fullEndpointId) {
	return extractPart(fullEndpointId, "endpoint:", null);

[Copilot]

Unnecessary ArrayList creation when currentShards is immediately populated with addAll(). Consider direct assignment: List currentShards = currentConfig.getShardsForNode(nodeId);

Suggested change

	List<String> currentShards = new ArrayList<>();
	// Get current shards (already in "redis:X" format)
	currentShards.addAll(currentConfig.getShardsForNode(nodeId));
	// Get current shards (already in "redis:X" format)
	List<String> currentShards = currentConfig.getShardsForNode(nodeId);

[Copilot]

String replacement logic is duplicated from RedisEnterpriseConfig.extractNumericShardId(). Consider using the existing utility method or creating a shared utility.

Suggested change

	mastersToFailover.add(shardId.replace("redis:", ""));
	mastersToFailover.add(RedisEnterpriseConfig.extractNumericShardId(shardId));

[Copilot]

The timeout is applied after repeatWhenEmpty, which means the timeout applies to the entire repeat sequence rather than individual checks. This could cause premature timeouts. Consider moving timeout inside the repeat or using takeUntil with a timer.

Suggested change

	}).repeatWhenEmpty(repeat -> repeat.delayElements(checkInterval)).timeout(timeout)
	}).repeatWhenEmpty(repeat -> repeat.delayElements(checkInterval))
	.takeUntil(Mono.delay(timeout))