因此，我将spring amqp用于以下配置：

ConnectionFactory factory = new ConnectionFactory();
factory.setAutomaticRecoveryEnabled(false);
CachingConnectionFactory ccf = new CachingConnectionFactory(factory);
ccf.setAddresses(addresses);
...

我正在使用amqp admin手动创建队列。在一些更新的库更新之前，使用此代码一切正常：

private void recreateContainer() {
    // if we are not already in recreation process
    if (this.recreatingContainer.compareAndSet(false, true)) {
      ExecutorService executor = Executors.newSingleThreadExecutor();
      executor.execute(
          () -> {
            int attempt = 0;
            while (this.listenerContainer.isRunning() && attempt < MAX_STOP_WAIT_ATTEMPTS) {
              try {
                attempt++;
                Thread.sleep(100);
              } catch (InterruptedException e) {
                Thread.currentThread().interrupt();
              }
            }
            if (attempt == MAX_STOP_WAIT_ATTEMPTS) {
              logger.error("Container took too long to stop");
            }
            logger.info(String.format("Recreating listener container after %s attempts", attempt));
            startNewContainer();
            this.recreatingContainer.set(false);
          });
    }
  }

  @Override
  public void onApplicationEvent(ListenerContainerConsumerFailedEvent event) {
    Object source = event.getSource();
    Assert.isInstanceOf(
        MessageListenerContainer.class,
        source,
        "source must be instance of MessageListenerContainer");
    MessageListenerContainer listenerContainer = (MessageListenerContainer) source;
    // if fatal and this state model is responsible for this listenerContainer
    if (event.isFatal() && this.listenerContainer == listenerContainer) {
      recreateContainer();
    }
  }

看来这是从 ListenerContainerConsumerFailedEvent 至 ListenerContainerConsumerTerminatedEvent 现在我的代码不工作了。
如果我更新了异常类型，队列将被重新创建，但我在停止listenercontainer时遇到问题。

ERROR [2021-03-08T16:52:10.291+01:00] lambda$recreateContainer$0: Container took too long to stop
INFO  [2021-03-08T16:52:10.291+01:00] lambda$recreateContainer$0: Recreating listener container after 100 attempts

现在，我得到了10个监听器，而不是5个监听器（因此，似乎监听器在重新创建队列后设法恢复，但在状态队列不存在时，我无法停止容器）。
任何帮助都将不胜感激。