[datadog_downtime_schedule] Add support for Downtime V2

[carlmartensen]

Implements the new datadog_downtime_schedule resource that wires the Downtime v2 API into the Datadog terraform provider. This resource is similar to the soon-to-be deprecated datadog_downtime except that it supports multiple recurrences and granular scope queries

Example:

resource "datadog_downtime_schedule" "downtime_schedule_example" {
  scope = "env:us9-prod7 AND team:test123"
  monitor_identifier {
    monitor_tags = ["test:123", "data:test"]
  }
  
  recurring_schedule {
    recurrence {
      duration = "1h"
      rrule    = "FREQ=DAILY;INTERVAL=1"
      start    = "2050-01-02T03:04:05"
    }
    recurrence {
      duration = "1d"
      rrule    = "FREQ=WEEKLY;INTERVAL=1"
      start    = "3050-01-02T03:04:05"
    }
    timezone = "America/New_York"
  }
  
  # one_time_schedule {
  #  start = null
  #  end = "2050-01-02T03:04:05Z"
  # }
  
  display_timezone                 = "America/New_York"
  message                          = "Message about the downtime"
  mute_first_recovery_notification = true
  notify_end_states                = ["alert", "warn"]
  notify_end_types                 = ["canceled", "expired"]
}

[Veske]

@carlmartensen
What is the recommendation for an indefinite downtime declaration when using this new API?

The provider requires us to specify at least one type of schedule:

Error: one_time_schedule or recurring_schedule must be set

We opted for specifying the one_time_schedule as one_time_schedule: {} but this is causing issues for recurrent runs of the same Terraform project with the following error:

Start time of downtimes in progress cannot be changed