hadoop - Can Spark store last 10 minutes data for real time application -


i trying make real time analytics application. overall procedure need follow is: logstash pumping .log data file kafka producer topic, kafka producer send kafka consumer topic. then, after need spark should read kafka consumer topic take input of last minutes data kafka consumer topic real time analytics , simultaneously kafka consumer should write data hdfs later historical analysis. have read spark can read hdfs can spark read directly kafka consumer topic? procedure following right or not? please suggest me.

yes. first of all, idea possible. however, might have consider performance of spark-stream, well. since spark-stream works in micro-batch environment, no in real-time system, spark stream can slow sometimes. if want better performance, better use flink.

see : flink-streaming


Comments

Popular posts from this blog

magento2 - Magento 2 admin grid add filter to collection -

Android volley - avoid multiple requests of the same kind to the server? -

Combining PHP Registration and Login into one class with multiple functions in one PHP file -