Introduction

Monitors the total message count (backlog) in a RabbitMQ queue during an experiment. If the backlog exceeds the configured threshold at any point during the check, the experiment fails.

This action targets queues, not nodes. For node-level monitoring (node down, alarms), use Check Node instead.

Use Cases

• Fail the experiment if the queue backlog grows beyond an acceptable threshold during a consumer fault injection.
• Verify that consumers keep up with message production while injecting network latency on consumer pods.
• Combine with Publish (Messages / s) to generate load and verify the backlog stays within bounds.

Parameters

Parameter	Description	Default
Duration	How long the check runs. The queue backlog is polled continuously for this duration.	30s
Acceptable Backlog	The maximum acceptable number of messages in the queue. If the total message count exceeds this value at any point during the check, the experiment fails. Specified as an integer (number of messages).	10