I am recently implementing a system that automatically replies to tweets that contain arbitrary hashtags. This system consists of a process that periodically crawls Twitter and a process that periodically replies to these tweets. Following the tradition of my company, these periodical jobs are implemented with working tables on RDMS that have a status column which would have values like "waiting", "processing" or "succeed". To ensure redundancy, I make multiple same processes running by leveraging low level locks.
My question is, I'm implementing periodically jobs with working tables in RDMS, how these jobs are implemented in generally.