enh(Zendesk): use the incremental endpoint for full sync

[aubin-tchoi]

Description

• Current state: there is an issue with Zendesk in that
  • we don't get all the tickets in the first full sync.
  • the incremental sync picks up 4 times more tickets in a month that the full sync.
• The Search API seems to be faulty: it contain a /count endpoint that returns a count of 3.8M tickets for the past 6 months for one connector, which seems unreasonable.
• This PR uses the incremental endpoint for the full sync, with a start date today - 6months and then filters the solved tickets.
• Since we can assume that most old tickets are solved, fetching all the tickets and filtering does not seem much more costly than using an unreliable endpoint that filters on fetch.
• Using the same endpoint for the full sync and the incremental sync seems very reasonable.
• Additionally, closed tickets are now also synced, solved tickets are automatically moved to closed in Zendesk.

Risk

Low, tested locally.

Deploy Plan

• Deploy connectors.

[fontanierh]

LGTM