develop
Fixed a race condition in the job service where a job could be started while the job service is stopping, producing a fault that retries the job on another instance when available.
Moved finalization of previous job attempts for recurring jobs to the schedule time vs the run time to avoid cleanup issues with message ordering / saga conflicts. Also adjusted partitioning to be by JobId instead of AttemptId for endpoint partitions.
Refactored job service for container-based configuration and eliminated the need for service instance configuration. Added extensions for Cosmos, Redis, MongoDB, Azure Table, EF Core
Add job type name to help diagnose problems.