eG Monitoring
 

Measures reported by MsSqlLogFileActTest

Every database created on an Microsoft SQL server is associated with a database file (.mdf) and a log data file (.ldf). All transactions to the database are logged in the log data file. The server will use the logged transactional information to restore the database after a crash, or when a transaction runs into issues and may have to be rolled back.

This test periodically monitors the I/O activity on and the growth in the size of the log file associated with each database. Using the metrics reported by this test, you can proactively detect bottlenecks while reading from or writing to the log file and excessive disk space usage by the log file.

The measures reported by this test are as follows:

Measurement Description Measurement Unit Interpretation
Num_of_writes Indicates the rate at which writes occurred on this log file. Write/Sec The detailed diagnosis of this measure provides the query that caused the waits; you may want to fine-tune the queries to reduce the number of waits.
Num_of_bytes_written Indicates the rate at which data was written to this log file. Kb/Sec Ideally, the value for this measure should be low. If the value for this measure is high, use the detailed diagnosis of the
Io_stall_write Indicates the total time taken to write to this logfile. Milliseconds/write A high value for this measure could indicate a bottleneck while writing to the log file.By comparing the value of this measure across log files, you can identify the log file to which write operations are taking too long to complete.
Num_of_reads Indicates the rate of reads from this log file. Read/Sec  
Num_of_bytes_read Indicates the rate at which data was read from this log file. Kb/Sec  
Io_stall_read Indicates the time taken to read from this log file. Milliseconds/read A high value for this measure could indicate a bottleneck while reading from the log file.

By comparing the value of this measure across log files, you can identify the log file to which read operations are taking too long to complete.
Io_stall Indicates the total time taken for I/O to complete on this log file. Milliseconds/IOP A high value for this measure could indicate an I/O bottleneck on this log file.
Size_on_disk Indicates the total size on disk of each logfile. MB This measure is used to determine the growth of the logfile.