chore: avoid blocking the whole waku node when retention policy is being applied

[Ivansete-status]

Background

The retention policy logic makes a very first check when the nwaku node is starting. However, the waitFor statement is being used for such initial message reduction, which therefore makes the whole node get blocked until the removal is completed.

That happens in

nwaku/waku/node/waku_node.nim

Line 765 in aeb77a3

let retPolRes = waitFor node.wakuArchive.executeMessageRetentionPolicy()

Note: Aside from that, I noticed that, in a database with 16Million rows and a retention policy of 12Milion rows, the actual execution got stuck the whole night, it is worth checking whether the rows-removal works well. That happened in Postgres.

Acceptance criteria

The application of the retention policy should never block the nwaku node.

[Ivansete-status]

Weekly Update

• achieved: avoid blocking the whole waku node when the retention policy is being applied