Home › Java exceptions › RejectedExecutionException

What it looks like

java.util.concurrent.RejectedExecutionException: Task ... rejected from java.util.concurrent.ThreadPoolExecutor[Running, pool size = 10, active threads = 10, queued tasks = 100, completed tasks = 5321]

Extends RuntimeException in java.util.concurrent.

What it means

A ThreadPoolExecutor applies its rejection policy when it cannot accept a task: either it has been shut down, or all threads are busy and the bounded queue is full. The default policy, AbortPolicy, throws RejectedExecutionException. The message prints the pool state, which tells you which of the two situations you are in: "Terminated" or "Shutting down" versus "Running" with a full queue.

Common causes

• Submitting work after shutdown(), often during application shutdown.
• A producer faster than the consumers, filling a bounded queue.
• Too few threads for the workload, or tasks that block for long.
• A pool created per request and never shut down, or a misconfigured Spring ThreadPoolTaskExecutor.

How to fix it

1. If the pool is shut down, stop submitting during shutdown and let in-flight work finish.
2. If the queue is full, apply backpressure: slow the producer, use CallerRunsPolicy, or size the pool and queue from measured load.
3. Shorten tasks or move blocking calls to a separate pool.
4. Monitor queue size and active threads so you see saturation before it rejects work.

Broken and fixed

ExecutorService ex = Executors.newFixedThreadPool(4);
ex.shutdown();
ex.submit(task); // rejected

ThreadPoolExecutor ex = new ThreadPoolExecutor(4, 4, 0, TimeUnit.SECONDS,
        new ArrayBlockingQueue<>(100),
        new ThreadPoolExecutor.CallerRunsPolicy());

🔎 Unravel the failing code 🔎 Unravel the fix

In Spring and enterprise code

Spring's @Async uses a task executor with an unbounded queue by default in Boot 3.x unless configured; setting spring.task.execution.pool.queue-capacity introduces rejection.

Debug it faster

Paste the whole stack trace into the stack trace beautifier to hide framework frames, jump to the root cause and find the first line of your own code. It runs in your browser, so the trace never leaves your machine.

Related exceptions