Distinguish between actual errors and when there are no tasks available

[kellpossible]

The contributors and verifiers receive try_lock errors when there are no tasks currently available for them to work on. This clutters the logs because these are not really errors, they are totally expected, and when an actual error occurs it gets lost in the noise.

[ibaryshnikov]

It is fixed for verifiers in #363, the endpoint replies with an Option<AssignedTask> type. Still and issue for contributors