Improve Host collection for the Pinger

[cortze]

Motivation

Due to the lack of multiple concurrent calls to a single host by the dht internal messenger (#805) and the large range of CID that we aim to track using the CID hoarder, the only way to reduce the impact of the time measurements that I came up with was to scale horizontally the number of hosts that make the +10k request.

Description

This PR presents the upgrade of the Pinger logic to:

1. Simplify it
2. implement a "smart" enough pool of hosts that will balance the pinging workload of a Continuous Study
   • Prioritized "idle" hosts
   • or host that is currently pinging the farthest CIDs (XOR distance) among the closest ones of the host pool
3. Fix proper routine closure
4. Fix debugging of DB batching
5. format code
6. add conf option for OP (experimental)