Tunning

• CBoard currently orientation to be view layer tool to handles and aggregate data of small size result set. Although CB has a flexible data source extension architecture, do not directly load deail data from any data products from all the details
• The first cost time of data loading is determined by the performance of the query. Please try to optimize your query or use Kylin, HBase, Redis, ES and other big data products as a data source
• As much as possible to control the size of dataset and query. We stress test down, 100W data using SELECT * FROM TABLE from the back-end to the front-end takes about 30s time, the browser uses about 100M memory, and then olap time within 5s, definitely not recommend this usage, please try to control the back-end Return data within 30W
• CBoard comes with performance optimization mechanism
  • In the absence of data sets under the premise of the same query in the same board will only load only once
  • In the case of pre-defined data set, the same data set query background once, custom aggregation expression and dimension data filtering can simplify the design, reduce the effect of the query
  • Query cache, board loaded data in the server cache for 12 hours to avoid the duplicate query