dromara/dynamic-tp
Dynamic Tp
Dynamic Thread Pool 是一个用于线程池动态管理、可观测性和编排的 Java 框架。它允许通过外部配置中心实时调整线程池参数，而无需重启应用程序。
该项目通过中间件编排脱颖而出，能够拦截并调优嵌入在第三方 Web 服务器和 RPC 框架中的线程池。它进一步支持分布式追踪，通过将执行上下文和元数据从提交线程传播到工作线程，以维持跨异步边界的可观测性。
该框架包含一个全面的监控和报警系统，可收集性能指标并通过多渠道分发触发通知。它具有基于阈值的报警引擎，带有滑动窗口抑制功能以防止报警风暴，以及用于监控执行超时的定时轮机制。
系统集成了广泛的中间件和基础设施，包括 Tomcat、Jetty、Undertow、Dubbo、gRPC、Hystrix 以及 RabbitMQ 和 RocketMQ 等消息队列。
Features
- Dynamic Thread Pool Adjusters - Provides tools for modifying thread pool parameters in real-time via external configuration centers.
- Concurrency Observability Tools - Provides a framework for exporting thread pool metrics and propagating tracing context across asynchronous execution boundaries.
- Thread Pool Context Propagators - Provides mechanisms for propagating thread-local context and tracing identifiers from submitting threads to worker threads.
- Dynamic Configuration Updates - Integrates with external configuration stores to enable real-time parameter updates without restarting the application.
- Custom Thread Pool Configurations - Enables the creation and definition of thread pool instances via code or configuration centers.
- Dynamic Thread Pool Proxies - Implements interceptors to apply runtime parameter changes and collect metrics from third-party thread pools.
- Dynamic Thread Pool Managers - Provides centralized management and runtime adjustment of thread pool parameters without application restarts.
- Middleware Pool Orchestration - Applies dynamic tuning and monitoring to thread pools embedded within third-party web servers and RPC frameworks.
- Configuration Store Integrations - Syncs runtime parameters with external stores like Nacos, Apollo, or Zookeeper to manage concurrent task execution.
- Runtime Parameter Injection - Allows real-time adjustment of thread pool core size and queue capacity via external configuration centers.
- Alert Thresholds - Monitors performance metrics against predefined numerical limits to trigger system notifications.
- Distributed Tracing Contexts - Propagates trace IDs and metadata across asynchronous boundaries to maintain distributed observability.
- Java Runtime Monitors - Collects real-time JVM performance metrics and triggers alerts on queue capacities and execution timeouts.
- Execution Timeout Monitoring - Tracks task queueing and processing duration using a time-wheel mechanism to detect execution timeouts.
- Execution Timeout Management - Tracks queueing and execution duration using a time-wheel mechanism to trigger alerts or interrupt threads.
- Dynamic Capacity Adjustment - Enables modifying the blocking queue size at runtime to handle fluctuating workloads.
- Thread Pool Anomaly Alerters - Triggers notifications when thread pool queue capacity or execution timeouts exceed configured thresholds.
- Duplicate Alert Suppression - Prevents alert storms by suppressing redundant notifications using a time-based sliding window.
- Thread Pool Metrics Collection - Gathers performance indicators via logging, JMX, or HTTP endpoints to track pool health and behavior.
