A simple RabbitMQ based work queue using a push architecture.
The work queue uses the RabbitMQ message broker to transmit messages from the server to the client. Each message is ACK'd by the client before being removed from the message broker.
The following is done for each message:
- Server broadcasts a message to
test_queuequeue - Client listens to the
test_queueand receives a message. - Client acknowledges the message, removing the message from the queue
To benchmark the performance of the message broker, the server broadcasts a set amount of equal-sized messages which are processed by the client.
The benchmark.py script starts the sender (server.py) and receiver (client.py) scripts in parallel and prints the time taken once both
scripts have finished sending and receiving the messages.
How to install (assuming Virtualenv for Python 3 is installed and RabbitMQ is running)
$ virtualenv -p python3 venv
$ source venv/bin/activate
$ pip install -r requirements.txt
Run a benchmark and print the time to send and receive the messages
$ # Send 5,000 messages, each with a size of 500 kB and measure the time it takes to send and receive the messages
$ python benchmark.py --payload-size=500000 --payload-count=5000
Run client and server individually
$ python server.py --payload-size=500000 --payload-count=5000
$ # In another terminal
$ python client.py --payload-count=5000