Bug 1024533

Summary: Metric chart's low and peak value is using wrong data (integer metrics return decimals)
Product: [JBoss] JBoss Operations Network Reporter: Larry O'Leary <loleary>
Component: Monitoring -- OtherAssignee: Mike Thompson <mithomps>
Status: CLOSED NOTABUG QA Contact: Mike Foley <mfoley>
Severity: urgent Docs Contact:
Priority: unspecified    
Version: JON 3.1.2CC: theute
Target Milestone: ---   
Target Release: JON 3.2.0   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2013-11-05 14:52:20 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On: 870561    
Bug Blocks:    
Attachments:
Description Flags
Metric graph page showing low value of 2 and peak of 3
none
Metric chart page showing low value of 2.05?? and peak of 2.95?? none

Description Larry O'Leary 2013-10-29 21:05:26 UTC
Description of problem:
When viewing the metric graph for a past period (such as from a couple weeks ago) we seem to be displaying the min and max values incorrectly for the specified period. This is very evident when working with metrics that are always of integer values such as an EAP 5 datasource's connection count. This metric can only be an integer. 

However, when reviewing data from a period from more then a couple weeks ago, it minimum and maximum values are displayed as fractions even though the data stored in the 1h, 6h, and 1d tables all clearly show that min and max values are 0 and 1. 

This actually may be a regression as this issue (or similar issue) had already been identified in bug 870561 and was supposedly fixed in JON 3.1.2. However, it is now obvious that this may not have been completely fixed.

Version-Release number of selected component (if applicable):
4.4.0.JON312GA

How reproducible:
Always

Steps to Reproduce:
1. Wait for the r00-r14, 1h, 6h, and 1d tables to be populated for the RHQ server's NoTxRHQDS total connections metric. 

   Please note that this will take 15 or more days. It may not be necessary to wait this long but the testing has revealed that this issue seems to happen when multiple metric tables are involved. It is not clear to which tables are actually involved.

2. View the RHQ server's NoTxRHQDS monitoring graphs page.,
3. Set the view range to be a week or so in the past and to be for an 8 hour period in where the metric was collected once every 20 minutes.
4. Take note of the low and peak values for the total connections graph.
5. View the RHQ server's NoTxRHQDS monitoring tables page and take note of the total connections metric minimum and maximum values.
6. Select the table row for the total connections metric and click the "chart selected metrics" button.
7. Take note of the Peak line and Low line on the total connections metric chart.


Actual results:
Low and Peak line on metric graph are aligned with a decimal (non-integer) on the y axis of the chart. These values do not match that of low and minimum or peak and maximum values displayed on the metric graph and table respectively.

Expected results:
All values should be the same and be an integer value.

Comment 1 Larry O'Leary 2013-10-29 21:20:14 UTC
Created attachment 817192 [details]
Metric graph page showing low value of 2 and peak of 3

Comment 2 Larry O'Leary 2013-10-29 21:21:25 UTC
Created attachment 817193 [details]
Metric chart page showing low value of 2.05?? and peak of 2.95??