v0.2.0

This release implements logic to address failure scenarios related to transient network or Redis errors. Previously, if a consumer failed to advance a job to its next state but then successfully registered a heartbeart, it was possible for it to retain ownership of that job indefinitely. Now, if a consumer encounters an error when trying to retry, kill, or acknowledge a job, it will immediately abort in order to ensure the job can be transferred to another consumer in a timely fashion.

This release also exposes a configurable logger interface and fixes some concurrency-related bugs.