Flume series: Record a large amount of Debezium data consumed at one time. The data contains relatively large dml statements, which causes the data to accumulate in the channel.
Flume series: Record a large amount of Debezium data consumed at one time. The data contains relatively large dml statements, which causes the data to accumulate in the channel.
Debezium collects a large amount of data, the data contains DMl statements, and the data itself is relatively large.
Flume consumes a large amount of Kafka topic data, and the data is accumulated in the channel, resulting in a relatively late time for the data sink to HDFS.
At this time, the data of the delay detection indicator org_apache_flume_channel_channel1_channelfillpercentage is relatively large, reaching more than 80.
Now it is necessary to quickly digest the accumulated data to ensure the timeliness of the data