Automating backup to S3

[novoj]

Currently all backups are "manual" - or require external API calls. This feature should allow configuring scheduled backups in regular intervals / or certain amout of "work". This rules are maintained on "engine" level API / state and support "wildcard" settings - e.g. * all catalogs are backed up each 24hour or after 500k transactions.

Task handling in a cluster is always "managed" by the leader, but might be executed by any of the replicas. Leader checks, that the task finished succesfully on the replica, if not it re-executes the task on different replica, or executes itself.