As a maintainer of the streaming updater I would like the pipeline to automatically restart when a transient error occurs so that I don't have to manually restart it.
Disabling automatic restarts was useful for debugging but now that the pipeline is covering all the usecases it might be better to enable a restart strategy.
- decide on one or multiple restart strategy
- the strategy should be configurable via options/property file
- the pipeline should restart itself on transient errors
- the pipeline should die on persistent errors