# InnoDB Change Buffer graph inaccurate? **URL:** \ **Category:** PMM 1.x\ **Created:** [January 18, 2018, 11:48pm UTC](https://forums.percona.com/t/innodb-change-buffer-graph-inaccurate/6137 "2018-01-18T23:48:30Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![bmorton](https://avatars.discourse-cdn.com/v4/letter/b/50afbb/32.png) [@bmorton](https://forums.percona.com/u/bmorton)\ **Post date:** [January 18, 2018, 11:48pm UTC](https://forums.percona.com/t/innodb-change-buffer-graph-inaccurate/6137/1 "2018-01-18T23:48:30Z")
I have a 50GB buffer pool and innodb\_change\_buffer\_max\_size set to 1. This should be 1% of 50GB, or about 500MB. innodb\_ibuf\_size reflects 1 (not sure what the unit is), but the graph shows 1.25GB utilized for this feature. I’m wondering if the math to arrive at the graphed number isn’t actually correct? Some info for diagnostics: mysql\> show status like ‘innodb\_ibuf%’; ±-----------------------------------±------+ | Variable\_name | Value | ±-----------------------------------±------+ | Innodb\_ibuf\_discarded\_delete\_marks | 0 | | Innodb\_ibuf\_discarded\_deletes | 0 | | Innodb\_ibuf\_discarded\_inserts | 0 | | Innodb\_ibuf\_free\_list | 82011 | | Innodb\_ibuf\_merged\_delete\_marks | 101 | | Innodb\_ibuf\_merged\_deletes | 9 | | Innodb\_ibuf\_merged\_inserts | 2219 | | Innodb\_ibuf\_merges | 1557 | | Innodb\_ibuf\_segment\_size | 82013 | | Innodb\_ibuf\_size | 1 | ±-----------------------------------±------+ 10 rows in set (0.00 sec) [url][grafana-dashboards/MySQL\_InnoDB\_Metrics.json at pmm-1.x · percona/grafana-dashboards · GitHub](https://github.com/percona/grafana-dashboards/blob/master/dashboards/MySQL_InnoDB_Metrics.json#L1717%5B/url%5D) ---
**Author:** ![Peter](https://sea1.discourse-cdn.com/flex019/user_avatar/forums.percona.com/peter/32/2_2.png) [@Peter](https://forums.percona.com/u/Peter)\ **Post date:** [January 29, 2018, 8:42pm UTC](https://forums.percona.com/t/innodb-change-buffer-graph-inaccurate/6137/2 "2018-01-29T20:42:53Z")
In your case you have “ibuf free list” of 82011 which is amount of space actually allocated in Insert buffer which is where 1.25GB come from the ibuf\_segment\_size is 82013 This graph does not make a difference between what is allocated in memory vs what is allocated in segment on disk We’ll think how this can be improved.