I am using a Spring Boot application with multiple instances and leveraging the RDelayedQueue from the Redisson library. I noticed that the library uses long timeout = System.currentTimeMillis() + delayInMs as the score for elements in a sorted set.
Here’s my concern, two tasks are offered to the delayed queue at the same time, resulting in the same timeout score.

• Instance 1: delayed.offer(“A”, delayTime, TimeUnit.MINUTES)
• Instance 2: delayed.offer(“B”, delayTime, TimeUnit.MINUTES)

Given that Redis is single-threaded and processes commands sequentially, my questions are:

1. How does Redisson handle the insertion and retrieval of elements with the same score in this scenario?
2. If the insertion of task “B” is processed first, will task “A” still be polled first because of its lexicographical order?
3. Does this behavior deviate from the traditional FIFO queue concept, and if so, how does RDelayedQueue ensure consistent polling order?

I would appreciate insights on how Redis manages this and any potential implications for using RDelayedQueue in a distributed environment with multiple instances.