opengoofy/hippo4j

Hippo4J is a Java framework for managing thread pools in running applications without restarting them or changing your code. A thread pool is a group of worker threads an application keeps ready to handle tasks, and tuning it correctly matters a lot under heavy load. Hippo4J lets you adjust thread pool settings while the application is live, watch what those pools are doing in real time, and receive alerts when something looks wrong. The typical pain points Hippo4J addresses are listed in its README: pools defined with the wrong size cause server overload, parameters that seemed fine in development become too small as traffic grows, tasks pile up and trigger rejection policies that drop work silently, and when an outage happens, there is no record of whether thread exhaustion was involved. Hippo4J adds visibility and control to solve all of these. The project supports two deployment modes. One depends on a configuration center (such as Nacos) that your team already runs, and the other is a standalone Hippo4J server you deploy yourself with no extra dependencies. Either way, there is a web console where you can see thread pool metrics, view charts of the last 30 minutes of activity, and change settings like core thread count, maximum thread count, or queue capacity on the fly. Alerts fire on four conditions: pool active rate (threads all busy), queue fill level, rejection policy triggers, and task execution time exceeding a threshold. Notifications can go to WeChat Work, DingTalk, Feishu, or email. Beyond the standard JDK thread pools in your own code, Hippo4J can also manage the thread pools inside frameworks like Dubbo, RabbitMQ, and RocketMQ, as well as the worker threads inside web containers like Tomcat, Jetty, and Undertow. The project is written in Java and is documented in Chinese with an English README also available.