# Reading MySQL Slow Query graph on Grafana

**URL:** https://forums.percona.com/t/reading-mysql-slow-query-graph-on-grafana/7336
**Category:** Other MySQL® Questions
**Created:** [December 9, 2019, 9:56am UTC](https://forums.percona.com/t/reading-mysql-slow-query-graph-on-grafana/7336 "2019-12-09T09:56:35Z")
**Posts on this page:** 2
**Page:** 1

<div class="post-metadata">

### Author: ![sh123](https://avatars.discourse-cdn.com/v4/letter/s/a9a28c/32.png) [@sh123](https://forums.percona.com/u/sh123)
#### Post date: [December 9, 2019, 9:56am UTC](https://forums.percona.com/t/reading-mysql-slow-query-graph-on-grafana/7336/1 "2019-12-09T09:56:35Z")

</div>

I’m having troubles understanding what the graph shows for the MySQL Slow Query graph, particularly with what the y axis represents.  
I’ve attached photos of some results I see, I understand the graph shows all slow queries that took longer than the long\_query\_time variable set. For my case the value is 0.25 . So I’m seeing all queries that took longer than 0.25s. But I thought I would be seeing number i.e. integers in the y axis to show number x slow queries within a specific time frame. But i’m seeing decimals, lower than 0.25 value which confuses me.

 ![slow3.PNG](https://us1.discourse-cdn.com/flex019/uploads/percona1/original/2X/1/125d32ab93a0a2ffe7879f0dc00b2524a206fd64.png)

 ![slow1.PNG](https://us1.discourse-cdn.com/flex019/uploads/percona1/original/2X/b/b9c8e23bd6a7e091deffb1f1c2ba1b4cceb7f554.png)

 ![slow2.PNG](https://us1.discourse-cdn.com/flex019/uploads/percona1/original/2X/f/f0db93ce97528f5fe90bbe05540e1ef23e335359.png)

---

<div class="post-metadata">

### Author: ![Rick\_James](https://avatars.discourse-cdn.com/v4/letter/r/22d042/32.png) [@Rick\_James](https://forums.percona.com/u/Rick_James)
#### Post date: [March 22, 2020, 5:29pm UTC](https://forums.percona.com/t/reading-mysql-slow-query-graph-on-grafana/7336/2 "2020-03-22T17:29:39Z")

</div>

I suggest it is slow queries per second. When the x-axis is on the order of second(s), then integral values make sense. When the x-increment is 5 minutes, then I imagine it is sampling every 10 seconds (or summarizing down to that), hence a reading of 0.10 would represent 1 slow query in the last 10 seconds. Be sure to check long\_query\_time. It defaults to a virtually useless 10 seconds. Crank that down to perhaps 1 (or even some fractional value) to catch more than the egregiously long-running queries.