Bug 877267 - Metrics on rare machines
Summary: Metrics on rare machines
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Beaker
Classification: Retired
Component: reports
Version: 0.9
Hardware: Unspecified
OS: Unspecified
medium
unspecified
Target Milestone: ---
Assignee: beaker-dev-list
QA Contact:
URL:
Whiteboard: Measurements
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2012-11-16 05:47 UTC by Min Shin
Modified: 2018-02-06 00:41 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: Enhancement
Doc Text:
Clone Of:
Environment:
Last Closed: 2016-10-26 04:04:20 UTC
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Red Hat Bugzilla 500835 1 None None None 2021-01-20 06:05:38 UTC
Red Hat Bugzilla 824534 0 urgent CLOSED [Beaker] RFE Default for scheduler to prioritise systems > 1 CPU 2021-02-22 00:41:40 UTC

Internal Links: 500835 824534

Description Min Shin 2012-11-16 05:47:30 UTC
Description of problem:

User would like to have a utilisation report of each rare machine. e.g. how much of time is machine idle / running recipe

Motivation: Question of "how rare is this configuration?" allows to tweak scheduler based on rarity of machine. For example a recipe, which has no host requirements should not get some rare configuration, which is in high demand for other tests.

Comment 1 Min Shin 2012-11-22 03:24:15 UTC
ncoghlan: The metrics aspect here should fall naturally out of implementing FR-8. Perhaps an additional requirement about presenting an "aggregate" percentage across multiple machines for the FR-8 metrics.

FR-8: Machine stats (https://bugzilla.redhat.com/show_bug.cgi?id=500835)

Comment 2 Nick Coghlan 2012-12-05 02:01:14 UTC

*** This bug has been marked as a duplicate of bug 193142 ***

Comment 3 Nick Coghlan 2013-01-04 02:33:47 UTC
This was duped on to 193142 when the planned implementation strategy for that bug was to track state changes explicitly for individual machines.

However, the aggregate metrics (overall and per-arch) eventually switched to a "sampling profiler" model (i.e. what percentage of machines are in a given state at each point in time) rather than tracking state changes explicitly on a per-machine basis.

Unfortunately, that approach does *not* address this feature request.

Comment 4 Dan Callaghan 2013-02-07 03:01:13 UTC
(In reply to comment #0)
> Description of problem:
> 
> User would like to have a utilisation report of each rare machine. e.g. how
> much of time is machine idle / running recipe

How do we know what is a "rare machine"? Is the point of this bug to give utilisation report for specific systems? (That is easy, we have code for it already.) Or is the point of this bug to find a way to do determine how rare a system is?

> Motivation: Question of "how rare is this configuration?" allows to tweak
> scheduler based on rarity of machine. For example a recipe, which has no
> host requirements should not get some rare configuration, which is in high
> demand for other tests.

This motivation suggests that the point of this report is actually to find rare machines.

Comment 8 Roman Joost 2016-10-26 04:04:20 UTC
At this point we don't think it's a good use of time to implement another report for the utilisation of rare machines. Getting to know certain utilisation metrics is important, but perhaps achieved in a different form that an additional report.


Note You need to log in before you can comment on or make changes to this bug.