Netflix/Hystrix · warning · RuntimeException

Interrupted while waiting for thread-pools to terminate. Poo

Error message

Interrupted while waiting for thread-pools to terminate. Pools may not be correctly shutdown or cleared.

What it means

HystrixThreadPool.shutdown(timeout, unit) stops every pool and awaits termination; if a waiting thread is interrupted during awaitTermination, it throws RuntimeException('Interrupted while waiting for thread-pools to terminate. Pools may not be correctly shutdown or cleared.') and the static threadPools map is left uncleared.

Source

Thrown at hystrix-core/src/main/java/com/netflix/hystrix/HystrixThreadPool.java:151

        }

        /**
         * Initiate the shutdown of all {@link HystrixThreadPool} instances and wait up to the given time on each pool to complete.
         * <p>
         * NOTE: This is NOT thread-safe if HystrixCommands are concurrently being executed
         * and causing thread-pools to initialize while also trying to shutdown.
         * </p>
         */
        /* package */static synchronized void shutdown(long timeout, TimeUnit unit) {
            for (HystrixThreadPool pool : threadPools.values()) {
                pool.getExecutor().shutdown();
            }
            for (HystrixThreadPool pool : threadPools.values()) {
                try {
                    while (! pool.getExecutor().awaitTermination(timeout, unit)) {
                    }
                } catch (InterruptedException e) {
                    throw new RuntimeException("Interrupted while waiting for thread-pools to terminate. Pools may not be correctly shutdown or cleared.", e);
                }
            }
            threadPools.clear();
        }
    }

    /**
     * @ExcludeFromJavadoc
     * @ThreadSafe
     */
    /* package */static class HystrixThreadPoolDefault implements HystrixThreadPool {
        private static final Logger logger = LoggerFactory.getLogger(HystrixThreadPoolDefault.class);

        private final HystrixThreadPoolProperties properties;
        private final BlockingQueue<Runnable> queue;
        private final ThreadPoolExecutor threadPool;
        private final HystrixThreadPoolMetrics metrics;
        private final int queueSize;

View on GitHub (pinned to 5ce3bc58c3)

Solutions

Restore the interrupt and retry shutdown: catch the RuntimeException, call Thread.currentThread().interrupt() is already-signal; re-invoke Hystrix.shutdown() with a longer/second attempt
Perform shutdown from a dedicated, non-interruptible cleanup path (e.g. contextDestroyed without container-imposed timeouts)
Increase the timeout passed to shutdown so awaitTermination finishes before any container deadline
Accept pooled-thread leakage at JVM exit (harmless when the process ends anyway) — only fix if the container keeps running

Example fix

// before
Runtime.getRuntime().addShutdownHook(new Thread(() -> Hystrix.shutdown()));
// after
Runtime.getRuntime().addShutdownHook(new Thread(() -> {
  try { Hystrix.reset(5, TimeUnit.SECONDS); }
  catch (RuntimeException e) { Thread.currentThread().interrupt(); /* JVM exiting */ }
}));

Defensive patterns

Strategy: try-catch

Validate before calling

null

Type guard

null

Try / catch

catch (RuntimeException e) { if (e.getMessage().startsWith("Interrupted while waiting for thread-pools")) { Thread.currentThread().interrupt(); /* optional: retry Hystrix.reset once with a longer timeout */ } }

Prevention

Call Hystrix.reset()/shutdown() from a dedicated cleanup thread with adequate timeout
Drain in-flight commands before shutdown
Never rely on interrupt-driven container timeouts for Hystrix teardown

When it happens

Trigger: Calling Hystrix.shutdown() / HystrixThreadPool.shutdown() and having the calling thread receive an interrupt (shutdown hook racing JVM exit, another thread interrupting, container thread interruption during redeploy).

Common situations: Application shutdown hooks where Hystrix shutdown races with JVM teardown; servlet container hot redeploys interrupting cleanup threads; tests that time out and interrupt the thread running Hystrix shutdown.

Related errors

AI-assisted analysis of Netflix/Hystrix@5ce3bc58c3 (2026-08-14). Data as JSON: /api/errors/d600dbe2bdd5ff47. Report an issue: GitHub.