DynamicTp Documentation

DynamicTp is a lightweight dynamic thread pool library that extends ThreadPoolExecutor to provide:

• Runtime Parameter Modification: Adjust core/max pool size, keep-alive time, etc., without redeployment.
• Monitoring & Alerting: Real-time status monitoring (20+ metrics) and notifications for config changes, thread activity, queue capacity, and rejections via WeCom, DingTalk, Feishu, Email, etc.
• Task Enhancement: Implement TaskWrapper for context propagation (e.g., MDC, TTL, OpenTelemetry).
• Middleware Management: Integrated support for Tomcat, Jetty, Dubbo, RocketMQ, gRPC, and more.
• Extensibility: SPI-based architecture for custom configuration centers, alerting, metrics, and task wrapping.

To integrate DynamicTp into your application, follow these four steps:

1. Add the dependency: Include the dependency for your chosen configuration center (e.g., Nacos, Apollo, Zookeeper, etc.).
2. Configure thread pools: Define your thread pool instances within your configuration center.
3. Enable DynamicTp: Add the @EnableDynamicTp annotation to your main application class.
4. Inject the executor: Use @Resource or @Autowired to inject the thread pool, or retrieve it manually using DtpRegistry.getExecutor("name").

You can integrate DynamicTp into your Spring-based application in 4 simple steps:

1. Add Dependencies: Include the dependency for your chosen configuration center (e.g., Nacos, Apollo, etc.).
2. Configure Thread Pools: Define your thread pool instances in your configuration center using the required configuration format.
3. Enable DynamicTp: Add the @EnableDynamicTp annotation to your Spring Boot main application class.
4. Inject and Use: Use @Resource or @Autowired to inject the thread pool, or retrieve it via DtpRegistry.getExecutor("name").

For more detailed configuration schemas, refer to the official documentation at https://dynamictp.cn.

DynamicTp can trigger real-time alerts based on configured thresholds. Supported alert dimensions include:

• Configuration Changes: Notifies when thread pool parameters are modified.
• Liveness (Activity) Alarms: Alerts when the thread pool is not functioning correctly.
• Queue Capacity Threshold Alarms: Alerts when the task queue reaches a certain limit.
• Rejection Alarms: Alerts when the rejection policy is triggered.
• Task Timeout Alarms: Alerts when tasks exceed execution or waiting time thresholds.

Supported notification channels include WeChat Work (企微), DingTalk (钉钉), Feishu (飞书), Email, and Yunzhijia (云之家). Custom notification logic can be implemented via SPI.

DynamicTp supports four metric collection modes. You can specify which modes to use via the collectorTypes configuration key. The default mode is Micrometer.

Available modes:

• Logging: Outputs metrics as JSON to a dedicated log file.
• Internal_logging: Outputs metrics as JSON to the application log.
• Micrometer: Uses the Micrometer facade to collect metrics for platforms like Prometheus or InfluxDB.
• Endpoint: Exposes a Spring Boot Endpoint for real-time metric retrieval via HTTP.

DynamicTp supports four types of monitoring data collection via the collectorTypes attribute. You can configure which type to use based on your observability stack:

• Logging: Outputs thread pool metrics in JSON format to a specified log file.
• Internal_logging: Outputs thread pool metrics in JSON format to the project's standard log files.
• Micrometer: Uses the Micrometer facade to collect metrics and export them to platforms like Prometheus or InfluxDB (requires Micrometer dependencies).
• Endpoint: Exposes an endpoint to retrieve the latest metrics via HTTP requests.

By default, Micrometer is used.

DynamicTp provides several specialized executor types to suit different business scenarios:

• DtpExecutor: The standard enhanced thread pool.
• EagerDtpExecutor: Optimized for IO-intensive scenarios.
• ScheduledDtpExecutor: For scheduled task execution.
• OrderedDtpExecutor: For scenarios requiring task ordering.