WaterSlide is a streaming event-at-a-time architecture for processing metadata. It is designed to take in a set of streaming events from multiple sources, process them through a set of modules ("kids"), and return meaningful outputs.
Other
113
stars
40
forks
source link
Bugfix for kafka and tuple2json kids, performance improvements, new largestring datatype #29
BUGFIX - fixed tuple2json kid to handle large complex objects
added largestring datatype (132KB) - requires complete recompile to use
added zerocopy options to kafka_out, better retries, and better tracking of failures
added time_in kid for triggering flushing on kids
added partition labeling for kafka_multi_in
added timeout adjustment when repeatedly polling sources and getting no data
BUGFIX - fixed tuple2json kid to handle large complex objects added largestring datatype (132KB) - requires complete recompile to use added zerocopy options to kafka_out, better retries, and better tracking of failures added time_in kid for triggering flushing on kids added partition labeling for kafka_multi_in added timeout adjustment when repeatedly polling sources and getting no data