SQL queue is growing in a peaktime

From VoIPmonitor.org
Jump to navigation Jump to search

Delay between active call and cdr view

The Active Calls view in the GUI displays the timestart of a calls(INVITEs) obtained from the VoIPmonitor sniffer service, whereas the CDR view shows the stored CDRs (after the call ends) from the database.

SQLq/SQLf

In the service status (expanded status, status line) under GUI -> Settings -> Sensors : status, the SQLq/SQLf values represent the size of the queue before the CDRs are pushed to the database. When SQLq/SQLf is high, it usually indicates that the database is unable to process requests in a timely manner, causing them to queue. Logging#SQLq/SQLf

Make sure that db config is not causing io overhead

If the MySQL configuration already follows the recommendations from the scaling section of our documentation (especially innodb_flush_log_at_trx_commit and innodb_buffer_pool_size):

https://www.voipmonitor.org/doc/Scaling#SSDs

More threads/connections to a db

you can also increase the number of threads used for connection to a db for particular use like CDRs - the VoIPmonitor sniffer service in /etc/voipmonitor.conf uses the option 

mysqlstore_max_threads_cdr = 4

However, if the database is waiting for storage I/O, increasing the number of threads will not help.

Retrieved from "https://www.voipmonitor.org//doc/index.php?title=SQL_queue_is_growing_in_a_peaktime&oldid=4951"