Hadoop Blueprints

Stream data analytics

Now let's start looking at the implementation of stream data analytics. Stream data analytics consists of two important elements:

• Loading streams of sensor data
• Data visualization using Grafana

Loading streams of sensor data

For batch data analytics we loaded data from Kafka to HDFS, but we will load streaming data into Open TSDB. To do this, first of all please make sure the following services are installed and tested successfully:

• Kafka
• Open TSDB
• Grafana

To extract the data from Kafka topics, we will be using Flume Kafka source and memory channel. But to load the data into Open TSDB, Flume does not provide a suitable sink by default, so I have written this simple sink. The code for the sink is available at https://github.com/deshpandetanmay/flink-opentsdb-sink ...