How to stop a RabbitMQ publisher missing heartbeats

In response to: https://stackoverflow.com/questions/73111224/connection-to-two-rabbitmq-servers

Running the project

Since this project depends on RabbitMQ, it's been set up to run two separate RabbitMQ instances side-by-side, using Docker and Docker Compose. The instances are called rabbitmq_upstream and rabbitmq_downstream, for the consumer to consume from and the publisher to publish to, respectively.

For the RabbitMQ admin UI, rabbitmq_upstream exposes port 7001 to localhost and rabbitmq_downstream exposes port 9001 to localhost. You can log in with the defaults of guest and guest.

You should be able to run the project with docker compose up -d.

NOTE: The consumer process will fail initially because the RabbitMQ containers take a while to get going.

Testing

The easiest way to see what's happening is to open up two terminals:

1. One with docker compose logs -f to tail the logs from the two RabbitMQ instances
2. The other with docker compose run --rm consumer to start the consumer-publisher process

The happy path

You'll want to convince yourself that messages are being fowarded correctly, which you can do by publishing a message to rabbitmq_upstream, which you can do at http://localhost:7001/#/queues/%2F/upstream_queue.

The same message should appear in the consumer logs as well as in rabbitmq_downstream at http://localhost:9001.

No messages about missed heartbeats should be appearing in the logs at any given time

Simulate no heartbeat handling

To simulate the situation where the publisher misses heartbeats, comment out the while loop in Publisher.run and uncomment the pass to make that method do nothing. Re-running the consumer with docker compose run --rm consumer should produce a missed heartbeat error in the RabbitMQ logs after about 20 seconds.

2022-08-01 12:32:21.442954+00:00 [error] <0.2784.0> missed heartbeats from client, timeout: 10s