rabbitmq AMQP::consume()

Question

AMQP function consume() is a blocking function with a callback, Is it possible to set a timeout for consume() function, so after specific amount of time it doesn't block anymore and the code execution completes ?

Endure

Answer 1

Yes, here's how:

$amqp = new AMQPConnection($your_connection_params);
$amqp->setTimeout($seconds);

Then when you call consume() on a queue, if no messages arrive within the timeout period, an AMQPException will be thrown from consume() with the message, "Resource temporarily unavailable". If you ever break out of consume() or hit a timeout, be sure to call cancel() on the queue object to properly reset the consumer. In order to do this, you need to generate a globally unique consumer tag and pass it in as an undocumented, third parameter to consume:

$tag = uniqid() . microtime(true);
$queue->consume($callback, $flags, $tag);
$queue->cancel($tag);

That way, you can call consume() again later without weird issues that will make your head spin.

Feminacy

FYI, I'm using the pecl amqp module, version 1.0.9. I tried upgrading to the latest version recently and my code completely broke. But the method I described above is working properly with amqp-1.0.9 and rabbitmq 3.1.3. – Feminacy

Ran into this scenario today and your answer lead me in the right direction. The setTimeout method has been deprecated in favour of the setReadTimeout. Also make sure to capture the AMQPConnectionException exception that consume throws when you timeout. AMQP version 1.4 – Sternmost