fix: archive - simplify and enhance async retention policy application

[Ivansete-status]

Description

This PR fixes an issue that made the whole wakunode2 to get stuck when applying the retention policy at the beginning of the execution. That happened because we used waitFor which blocks everything until the task end.

With that PR, we perform a better retention policy application approach, i.e. a fully asynchronous task that can be canceled anytime. We also simplify the approach by avoiding using a "timer".

Changes

• Use of async retention policy task

How to test

Test case 1

1. Start a node with Store configured, with _Postgres for example, and then it many messages, with a wide retention policy (capacity:10000000)
2. Restart the node but in this case with shorter retention policy, capacity:10.
3. The node applies the retention policy, without blocking any other task.

Test case 2

1. Start the node, changing the code a little so that the retention policy is applied every 3 seconds.
2. Keep sending/publishing messages to the node.
3. The amount of messages stored in the Postgres database fluctuates around the retention policy value, considering that "capacity" is being used as a retention policy.

Issue

closes #2238

[github-actions]

You can find the image built from this PR at

quay.io/wakuorg/nwaku-pr:2278

Built from cde6c6b