rptest: Fix race condition in CloudRetentionTest

[Lazin]

Fixes #15893

The test has race condition. It starts consumer before all data is produced and before retention kicks in. The original intention was to let the consumer validate the data after retention. Basically, to check that the data is readable. But because of that the consumer were reading data that was removed by retention. It was racing with the producer. Largely simplifed version of what was happening there.

• The producer produces offsets in range [0...1000].
• The retention removes offsets in range [0...500).
• The consumer is started concurrently with the producer with the expectation that it will consume offsets in range [500...1000]. We're validating the number of messages consumed.
• The consumer actually manages to consume offsets in range [0...600]. So the check based on number of messages passes.
• In CDT the consumer doesn't manage to consume enough data because of different performance characteristics. It consumes in range [0...100] and check fails.

This PR fixes this by starting the consumer after retention is confirmed.

Backports Required

• none - not a bug fix
• none - this is a backport
• none - issue does not exist in previous branches
• none - papercut/not impactful enough to backport
• v23.3.x
• v23.2.x
• v23.1.x

Release Notes

• none

[vbotbuildovich]

/backport v23.3.x