I've opened #13 to get more information as to what's going wrong, but running our CI is currently blocked by SR-15362. Once that's resolved (it has a PR open), i should be able to look into this more.

My current hunch is that there's either a resolution or precision issue with the systemUptime clock on one of the CI machines, that's causing an iteration to appear to have run either at or before the time that the previous one ended. Depending on the severity of it, it could be possible to change the assertion to allow for next.start >= result.end, or to add another delay after recording the end timestamp, which will still cause the test to fail if the queue were being run concurrently.

The extra logging revealed that some loop iterations had the same "start" timestamp as the previous "end" timestamp. I updated the assertion to allow this, and ran CI 10 times without an issue. The PR has been merged.