@Async TaskExecutor transition to bounded queue timeout handling

Task Execution Resource Management

Spring Framework's @Async functionality relies on a TaskExecutor to manage asynchronous workloads. While SimpleAsyncTaskExecutor is available, it does not support bounded queues, which can lead to resource exhaustion during high-load spikes.

p>Transitioning to ThreadPoolTaskExecutor allows for the definition of corePoolSize, maxPoolSize, and queueCapacity. However, when the queue capacity is reached, the default behavior triggers a RejectedExecutionException if a specific RejectedExecutionHandler is not configured.

Configuration Uncertainty

There is a design trade-off between using a CallerRunsPolicy to throttle the producer and implementing a custom timeout mechanism to ensure system responsiveness without crashing the application context.

• How does the choice of RejectedExecutionHandler impact the propagation of timeouts back to the original caller?
• What is the recommended approach for ensuring that CompletableFuture.cancel(true) reliably interrupts a task running within a bounded ThreadPoolTaskExecutor?