[Recovery] Add delay after which fail count is reset

[raulb]

Feature description

As described in this design document, to control intermittent errors, the recovery mechanism will have a parameter named HealthyAfter that

This pull-request implemented the restart of a pipeline to be recovered, but it was considered tackling HealthyAfter separately considering it'd increase its scope notably.

Some details

• The backoff will be reset only when a pipeline processes at least one record or runs without encountering an error for a predefined time frame.
• Each pipeline will control its own HealthyAfter duration https://github.com/ConduitIO/conduit/blob/raul/restart-with-recovery/pkg/lifecycle/service.go#L55.
• This will need to be updated once we have this delay in place https://github.com/ConduitIO/conduit/blob/raul/restart-with-recovery/pkg/lifecycle/service.go#L752-L756.

Possible solution

Spawn a goroutine that resets the backoff after a certain time.