keepAliveTime

keepAliveTime is a configuration parameter used by thread pool implementations to determine how long idle worker threads may remain alive before being terminated. It helps balance responsiveness and resource usage by allowing the pool to shrink during quiet periods while retaining a ready set of threads for bursts of work. The concept is tightly related to the core pool size and the maximum pool size.

In typical thread pool designs, there are core threads and non-core threads. Core threads are kept alive to handle incoming tasks, even when idle (unless core threads are allowed to time out). Non-core threads are created as needed to handle bursts and are eligible for termination after they have been idle for the keepAliveTime interval. If the pool has more than the core pool size, an idle non-core thread older than keepAliveTime will be terminated. If allowCoreThreadTimeOut is enabled, core threads can also time out and be terminated after the same interval.

The value of keepAliveTime is specified with a time unit, such as seconds or minutes. In Java's ThreadPoolExecutor, for example, keepAliveTime is configurable at construction and can be adjusted at runtime (typically via setKeepAliveTime) and is measured in a TimeUnit. The default is often 60 seconds for non-core threads.

Practical guidance: a shorter keepAliveTime reduces idle resource usage but may increase thread churn during bursts; a longer value reduces churn but retains more idle threads. Tuning should consider task duration, arrival rate, and desired latency. If core threads should also time out, enable allowCoreThreadTimeOut and set an appropriate keepAliveTime value.