It makes sense to revise that list and check what stats can be removed or reimplemented as INNODB_METRICS counters. The benefits of INNODB_METRICS over status variables is ability to enable/disable counters selectively, retrospection, stopwatch-like time interval values, etc. The downside is a counter-intuitive configuration, see http://bugs.mysql.com/bug.php?id=69099
Naturally, this should be implemented before the GA release, as changing the list of available status variables will be problematic afterwards.
Whiteboard
Implementing RO and RW trx list lengths in I_M tables belongs to this blueprint as well.
Waiting for some peformance numbers.
Current idea is not to remove the status vars (as they are already ported), and consider addding the I_M support on the top (InnoDB has provisions for that). Then I_M support decision needs not to happen before the GA.
It appears that this issue is no longer being worked on, so we are closing it for housekeeping purposes. If you believe the issue still exists, please open a new ticket after confirming it's present in the latest release.
Registered by Alexey Kopytov on 2013-05-19
There's some overlap between XtraDB extended status variables from http://www.percona.com/doc/percona-server/5.5/diagnostics/innodb_show_status.html and values in the INNODB_METRICS table in 5.6: http://dev.mysql.com/doc/refman/5.6/en/innodb-metrics-table.html
It makes sense to revise that list and check what stats can be removed or reimplemented as INNODB_METRICS counters. The benefits of INNODB_METRICS over status variables is ability to enable/disable counters selectively, retrospection, stopwatch-like time interval values, etc. The downside is a counter-intuitive configuration, see http://bugs.mysql.com/bug.php?id=69099
Naturally, this should be implemented before the GA release, as changing the list of available status variables will be problematic afterwards.
Whiteboard
Implementing RO and RW trx list lengths in I_M tables belongs to this blueprint as well.
Waiting for some peformance numbers.
Current idea is not to remove the status vars (as they are already ported), and consider addding the I_M support on the top (InnoDB has provisions for that). Then I_M support decision needs not to happen before the GA.
Note also the deadlock count (bug 1192052).
Linked LP bugs:
https://bugs.launchpad.net/percona-server/5.6/+bug/1182837
https://bugs.launchpad.net/percona-server/5.6/+bug/1192052