[automatic failover] Implement HealtStatusManager + weighted endpoints

[atakavci]

• weighted cluster seleciton
• Healtstatus manager with initial listener and registration logic
• pluggable health checker strategy introduced, these are draft NoOpStrategy, EchoStrategy, LagAwareStrategy,
• fix failing integration tests impacted by weighted clusters

[atakavci]

@a-TODO-rov, is this what you suggest? if not, could you provide code snippet on your suggestion?

Suggested change

	if (newStatus.isHealthy()) {
	if (clusterWithHealthChange.isFailbackEnabled() && activeCluster != clusterWithHealthChange) {
	// lets check if weighted switching is possible
	Map.Entry<Endpoint, Cluster> failbackCluster = findWeightedFailbackCluster();
	if (failbackCluster == clusterWithHealthChange
	&& clusterWithHealthChange.getWeight() > activeCluster.getWeight()) {
	setActiveCluster(clusterWithHealthChange, false);
	}
	}
	} else if (clusterWithHealthChange == activeCluster) {
	iterateActiveCluster();
	}
	if (clusterWithHealthChange == activeCluster && !newStatus.isHealthy()) {
	iterateActiveCluster();
	return;
	}
	if (newStatus.isHealthy() && clusterWithHealthChange.isFailbackEnabled() && activeCluster != clusterWithHealthChange) {
	// lets check if weighted switching is possible
	Map.Entry<Endpoint, Cluster> failbackCluster = findWeightedFailbackCluster();
	if (failbackCluster == clusterWithHealthChange
	&& clusterWithHealthChange.getWeight() > activeCluster.getWeight()) {
	setActiveCluster(clusterWithHealthChange, false);
	}
	}
	}

[ggivo]

Hi
Looks good in general.
Adding some comments and also have some concern around possible syncronisation issue when combining HealtStatusChecks with existing CB failover logic.

[atakavci]

hey @dengliming ,
what about these comments you had via email. Could you explain and if this is something not planned, please remove or disable the automation or whatever process this is.
This is kind of spamming our PR's , so thank you for your cooperation.

[Copilot]

Pull Request Overview

This PR refactors the multi-cluster failover provider to use weighted, endpoint-based selection and adds a pluggable health check framework.

• Introduce HealthStatusManager and health check strategies (Echo, LagAware) to monitor endpoint health.
• Refactor MultiClusterPooledConnectionProvider from index-based to weighted, endpoint-driven failover (iterateActiveCluster, setActiveCluster).
• Update tests and client APIs (UnifiedJedis, CommandObjects) to exercise the new health/failover behavior.

Reviewed Changes

Copilot reviewed 25 out of 25 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/test/java/redis/clients/jedis/scenario/ActiveActiveFailoverTest.java	Switch to setActiveCluster & endpoint-based cluster lookup
src/test/java/redis/clients/jedis/providers/MultiClusterPooledConnectionProviderTest.java	Update to weighted selection and iterateActiveCluster API
src/test/java/redis/clients/jedis/misc/AutomaticFailoverTest.java	Replace incrementActiveMultiClusterIndex with iterateActiveCluster
src/test/java/redis/clients/jedis/mcf/HealthCheckTest.java	Add unit tests for HealthStatusManager and health checks
src/test/java/redis/clients/jedis/mcf/HealthCheckIntegrationTest.java	Integration tests for disabling/default/custom health strategies
src/test/java/redis/clients/jedis/failover/FailoverIntegrationTest.java	Refactor integration tests for new builder options & failover API
src/main/java/redis/clients/jedis/providers/MultiClusterPooledConnectionProvider.java	Core refactor to weighted, endpoint-keyed clusters and health
src/main/java/redis/clients/jedis/mcf/RedisRestAPIHelper.java	New REST helper for external DB availability checks
src/main/java/redis/clients/jedis/mcf/LagAwareStrategy.java	Implement lag-aware health check strategy
src/main/java/redis/clients/jedis/mcf/HealthStatusManager.java	Central health status manager with listener support
src/main/java/redis/clients/jedis/mcf/HealthStatusListener.java	Interface for health status event listeners
src/main/java/redis/clients/jedis/mcf/HealthStatusChangeEvent.java	Event object for status changes
src/main/java/redis/clients/jedis/mcf/HealthStatus.java	Enum encapsulating healthy/unhealthy state
src/main/java/redis/clients/jedis/mcf/HealthCheckStrategy.java	Strategy interface for health checks
src/main/java/redis/clients/jedis/mcf/HealthCheckCollection.java	Thread-safe collection of ongoing health checks
src/main/java/redis/clients/jedis/mcf/HealthCheck.java	Schedules & executes periodic health checks
src/main/java/redis/clients/jedis/mcf/EchoStrategy.java	Echo-based health check strategy
src/main/java/redis/clients/jedis/mcf/CircuitBreakerFailoverBase.java	Update to use iterateActiveCluster for failover
src/main/java/redis/clients/jedis/mcf/CircuitBreakerCommandExecutor.java	Drop obsolete options and use cluster retry flags
src/main/java/redis/clients/jedis/UnifiedJedis.java	Adjust constructors and add echo command
src/main/java/redis/clients/jedis/MultiClusterClientConfig.java	Extend config API for weight, health checks, retry/failback
src/main/java/redis/clients/jedis/HostAndPort.java	Implement new Endpoint interface
src/main/java/redis/clients/jedis/CommandObjects.java	Add echo command object

Comments suppressed due to low confidence (4)

src/main/java/redis/clients/jedis/mcf/RedisRestAPIHelper.java:38

• [nitpick] There's a placeholder comment here but no actual logging of the IOException. Consider injecting a logger and recording the exception for troubleshooting.

        HttpURLConnection getConnection = createConnection(bdbsUri, "GET");

src/main/java/redis/clients/jedis/mcf/CircuitBreakerFailoverBase.java:41

• The comment references the old incrementActiveMultiClusterIndex() method. It should be updated to mention iterateActiveCluster() to match the current implementation.

        try {

src/test/java/redis/clients/jedis/mcf/HealthCheckTest.java:300

• [nitpick] The comment says the default is null, but the code asserts the default is EchoStrategy.DEFAULT. Please update the comment to reflect the actual default strategy.

        assertEquals(EchoStrategy.DEFAULT, clusterConfig.getHealthCheckStrategySupplier()); // Default is null (no health check)

src/main/java/redis/clients/jedis/mcf/CircuitBreakerFailoverBase.java:53

• [nitpick] This commented-out code references a non-existent method and clutters the code. Consider removing it for clarity.

                // int activeMultiClusterIndex = provider.incrementActiveMultiClusterIndex1();

[ggivo]

LGTM

Let's start merging them in the feature branch so we have a consistent state there