MySQL Bugs: #49214: Graphs for server health using relative scale

Bug #49214	Graphs for server health using relative scale
Submitted:	30 Nov 2009 16:38	Modified:	11 Dec 2009 17:52
Reporter:	Todd Farmer (OCA)	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Workbench	Severity:	S3 (Non-critical)
Version:	5.2 r4643	OS:	Windows (XP)
Assigned to:	Mike Lischke	CPU Architecture:	Any

Description:
There are two "Server Health" graphs which use a relative scale:  Connections and Traffic.  This results in both graphs having at least one graphed "100%" point for the short-duration (~1m) sampling size.  I don't know if there are any indications when such statistics approach critical thresholds (eg, max_connections) or not, but these graphs effectively suggest - visually - that the server is operating at near-maximum capacity in these areas.

How to repeat:
Look at server graphs.

Suggested fix:
1. Use max_connections as scale for connections.
2. Check with MEM team how they graph Traffic.  Maybe use stepped scale (1MB, 10MB, 100MB, 1GB - the first value larger than current max value).

Fixed in 5.2.

I didn't use max_connection, though, to set the upper limit, but instead applied the stepwise range adjustment, which was planned anway. The same solution was also chosen for the traffic graph (with different thresholds).

these two graphs use a variable scaling now.
change will go into 5.2.11

An entry has been added to the 5.2.11 changelog:

The server health graphs Connection Usage and Traffic, in the Server Status panel of Admin tab, appeared to indicate the server was operating at 100% capacity, even when this was not the case.

MySQL Workbench has been changed to use variable scaling, rather than linear scaling, for these graphs.

Hey everyone,

The graphs using a relative scale, specifically the Connections and Traffic graphs, seem to be misleading due to the presence of the "100%" points. It gives the impression that the server is operating at near-maximum capacity, which may not necessarily be the case.

To fix this, I have a couple of suggestions. Firstly, for the Connections graph, using the max_connections as the scale would provide a more accurate representation of the server's capacity. This way, we can better understand when statistics approach critical thresholds, ex. here: https://www.medtechdive.com/press-release/20230605-transforming-dialysis-care-andersen-unv...

Secondly, it would be beneficial to consult with the MEM team to determine how they graph the Traffic. Implementing a stepped scale, such as 1MB, 10MB, 100MB, and 1GB, with the first value larger than the current max value, could help improve the visualization and prevent misinterpretation.

I believe these solutions would address the issue and provide more reliable insights into the server health. Let's work together to resolve this bug and enhance the accuracy of our monitoring system.