chore(shard-distributor): increase observability of the leader election and the leader processor

[arzonus]

What changed?

• More logs have been added around runElection function in the leader election
• Changed the Log level in the leader processor
• The rebalancing loop metrics became namespace-related

Why?

• We have observed instability around the leader election process; however the current observability level didn't catch the exact reason for it.
• The current metrics are emitted per instance level, not per namespace. To be able to capture a stuck rebalancing loop per namespace, the metric should be emitted per namespace.

How did you test it?

• Unit tests

Potential risks

Release notes

Documentation Changes