Extending uGroup

Extending uGroup

uGroup uses interfaces to allow custom implementations. Define a Spring @Bean to override any default.

MetricsProvider

Why override:

• Integrate with your existing metrics system (Datadog, StatsD, CloudWatch, etc.)
• Add custom tags (environment, team, service) to all metrics
• Route metrics to multiple backends simultaneously

@Bean
public MetricsProvider metricsProvider() {
    return new DatadogMetricsProvider(datadogClient);
}

ClusterConfig

Why override:

• Dynamic cluster discovery from service registry (Consul, ZooKeeper, etc.)
• Multi-cluster support with automatic failover
• Load Kafka credentials from secrets manager (Vault, AWS Secrets Manager)
• Environment-specific configuration (dev/staging/prod)

@Bean
public ClusterConfig clusterConfig() {
    return new ConsulClusterDiscovery(consulClient);
}

WatchListProvider

Why override:

• Load consumer groups from a database or API
• Integrate with your deployment system to auto-discover consumers
• Filter based on custom metadata (team ownership, SLA tier, etc.)
• Sync with a configuration management system

@Bean
public List<WatchListProvider> watchListProviders() {
    return List.of(
        new DatabaseWatchListProvider(dataSource),  // From your metadata DB
        new KubernetesWatchListProvider(k8sClient)  // Auto-discover from K8s
    );
}

OffsetCache

Why override:

• Distributed caching with Redis/Memcached for multi-instance deployments
• Persistent cache that survives restarts
• Share offset state across uGroup instances for HA setups

@Bean
public OffsetCache offsetCache() {
    return new RedisOffsetCache(redisClient);
}

OffsetFetcher

Why override:

• Add caching layer for Kafka API calls
• Implement retry logic or circuit breakers
• Use a different Kafka client library

@Bean
public OffsetFetcher offsetFetcher() {
    return new CachedOffsetFetcher(new DirectOffsetFetcher(clusterConfig));
}