This topic describes how to use the traffic flow analysis feature. This feature allows you to view the information about the top 10 tables sorted by request count and the top 10 tables sorted by data volume. This feature also allows you to view the information about the tables whose regions are ranked top 10 by request count per node.

Procedure

  1. Choose Diagnose > Traffic Flow Analysis. On the page that appears, you can view the analysis results.
  2. After you select a zone, select a group and specify the snapshot time. Then, you can view the table traffic records based on the specified group and the specified snapshot time.
Note
  • TopN represents top 10.
  • A TopN table snapshot is used to filter data and display the top 10 records. These records contain the information about the top 10 tables sorted by request count and the top 10 tables sorted by data volume. These records also contain the information about the tables whose regions are ranked top 10 by request count per node. If a suspected hot spot exists, the details are highlighted in red.
  • The Realtime TopRegion diagnostic tool is used to identify real-time hot spots more accurately so that you can locate issues more accurately. For more information about the procedure, see View real-time hot-spot regions.

Introduction to three types of tables

Top 10 tables sorted by request count: Tables are sorted based on the totalRate and the records about the top 10 tables are displayed. The records contain the data of core performance metrics for the tables. If a suspected hot-spot region exists, a warning message in red appears at the top of the user interface. In addition, the system uses the red color to highlight the record that contains the information about the suspected hot-spot region. This helps you locate the suspected abnormal table and the corresponding node information in an easy and quick manner.

Note To view the information about the top 10 regions for each table, click the plus sign (+) at the beginning of each record for the table. The regions are sorted based on the same dimension as that for sorting tables. If a suspected hot-spot region exists, the corresponding details are highlighted in red.
Request count

Top 10 tables sorted by data volume: Tables are sorted based on the totalData and the records about the top 10 tables are displayed. The records contain the data of core performance metrics for the tables. If a suspected hot-spot region exists, a warning message in red appears at the top of the user interface. In addition, the system uses the red color to highlight the record that contains the information about the suspected hot-spot region. This helps you locate the suspected abnormal table and the corresponding node information in an easy and quick manner.

Note To view the information about the top 10 regions for each table, click the plus sign (+) at the beginning of each record for the table. The regions are sorted based on the same dimension as that for sorting tables. If a suspected hot-spot region exists, the corresponding details are highlighted in red.
Tables sorted by data volume

Tables whose regions are ranked top 10 by request count per node: Tables are sorted based on the request count per node and the records about the top 10 tables are displayed. The records contain the data of core performance metrics for the tables.

Note In some scenarios, the traffic volume of a single node is high, but the traffic volume of the table associated with the node is not ranked the top 10. The sorting results help you solve the issues in these scenarios.
Single node