fix: Better error logs in sharding mismatch and avoid checking clusterId if clusterId == 0

[Ivansete-status]

Description

1. Enhance error management inside isClusterMismatched
2. Avoid validation inside isClusterMismatched if the node is using clusterId == 0. This aims to allow the node to start successfully and avoid the appearance of "cluster id mismatch configured shards" and the quit(QuitFailure). In other words, the node couldn't start due to this proc even though the clusterId param was not being set.

[github-actions]

You can find the image built from this PR at

quay.io/wakuorg/nwaku-pr:2445

Built from 188e841

[SionoiS]

It was this way before. See here https://github.com/waku-org/nwaku/pull/2368/files

You have to check if shards have been set when cluster id == 0 too, otherwise no error is returned when a node is launched with only shards and no cluster id (default 0).

[gabrielmer]

This PR would also reintroduce #2311 , I think we should allow nodes to start without subscribed shards and subscribe later dynamically

[SionoiS]

allow nodes to start without subscribed shards and subscribe later dynamically

I would be very ergonomic indeed.

[Ivansete-status]

Thanks for the comments guys!
Disregard this PR.
For some reason, last week we had many restarts in waku.fleet due to this proc, and the PR was motivated to overcome that. It also happened to me locally but now it works well :)

Anyways, I'll close this.