GigaChannel community DB access batching

[ichorid]

Currently, for each incoming GigaChannel community gossip packet we open a separate DB session and process packets individually. This will really start to stress the DB when the gossip stream will rise to dozens of packets per second. We should solve this problem somehow, for example by batching DB access.

[synctext]

Note, IO Batching slowed us down in the past.

[ichorid]

That's why we're gonna "science this out" by setting up some meaningful experiments.

[synctext]

@kozlovsky Are we still closing and opening DB sessions for each incoming UDP packet?
Are we expecting any gains from refactoring this?

[drew2a]

It is not relevant anymore:

• Channels Retirement #7669