eG Monitoring
 

Measures reported by BobiDFSQueryTest

The efficiency of the Data Federation Service can be measured by the quality of the queries the data federation administration tool helps build. By periodically checking the feedback returned by the data federation query engine, administrators can determine the number of queries that may have to be optimized before being executed. The BobiDFSQueryTest test helps administrators perform this check at regular intervals.

With the help of the metrics reported by this test, administrators can quickly figure out if there are any queries that are consuming disk space and memory resources. In addition, the test also measures the query load on the data federation query engine, reports the count of queries in various stages of processing, and thus indicates where most queries are spending time – in analysis? In optimization? In execution? This way, the test points to bottlenecks in query processing and the source of these bottlenecks.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
noOfRunQry Indicates the number of running queries. Number The value of this measure includes both memory-consuming and non-consuming queries.
qryConsumingMem Indicates the number of queries consuming memory. Number A non-zero value for this measure indicates that one/more queries are consuming memory resources. Queries that are consuming memory excessively could be candidates for optimization. You may want to view the query plan of every query in the data federation administration tool to know how much memory each query is consuming, which operators are consuming maximum memory, and how the query can be restructured to consume less memory.
qryUsingDisk Indicates the number of queries using disk space. Number A non-zero value for this measure indicates that one/more queries are consuming disk space. Queries that are consuming disk space excessively could be candidates for optimization.
qryWaitResourcee Indicates the number of running queries currently waiting for execution. Number  
failedQueries Indicates the number of queries that failed. Number Ideally, the value of this measure should be 0. A non-zero value could indicate that one/more queries have failed. You can use the Query Monitoring tab of the data federation administration tool to view the exception that caused the query to fail and to troubleshoot the failure.
qryInQryAnalyzeStep Indicates the number of running queries currently in analyze step. Number If, at any given point in time, the number of queries being analyzed are way more than the count of queries being executed, it could mean that queries are spending too much time in the analysis step.
qryInQryOptimizeStep Indicates the number of running queries currently in optimization step. Number If, at any given point in time, the number of queries being optimized are way more than the count of queries being executed, it could mean that too many inefficient/unoptimized queries are running.
qryInQryExectionStep Indicates the number of running queries currently in the execution step. Number  
producedByQueryExe Indicates the amount of data output by the executed queries. MB  
TransfFrmDataSource Indicates the amount of data read from data sources. MB  
memUsedByQueryExe Indicates the total amount of memory used by running queries. MB A high value is indicative of high memory consumption by one/more queries. Queries that are consuming memory excessively could be candidates for optimization. You may want to view the query plan of every query in the data federation administration tool to know how much memory each query is consuming, which operators are consuming maximum memory, and how the query can be restructured to consume less memory.
diskUsedByQryExe Indicates the total amount of disk space used by running queries. MB A high value is indicative of high disk space usage by one/more queries. Queries that are consuming disk space excessively could be candidates for optimization.
memUsedByMetadataCache Indicates the amount of memory used for caching metadata, statistics and connectors configuration. MB