Bug 1464924 - [RFE] Expose the provider disable option in the UI
[RFE] Expose the provider disable option in the UI
Status: VERIFIED
Product: Red Hat CloudForms Management Engine
Classification: Red Hat
Component: UI - OPS (Show other bugs)
5.8.0
Unspecified Unspecified
unspecified Severity unspecified
: GA
: 5.9.0
Assigned To: Ari Zellner
Pavel Zagalsky
container
: FutureFeature, RFE
Depends On:
Blocks:
  Show dependency treegraph
 
Reported: 2017-06-26 05:00 EDT by Ari Zellner
Modified: 2017-11-30 05:39 EST (History)
12 users (show)

See Also:
Fixed In Version: 5.9.0.1
Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of:
Environment:
Last Closed:
Type: Bug
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: Container Management


Attachments (Terms of Use)

  None (edit)
Description Ari Zellner 2017-06-26 05:00:28 EDT
Description of problem:
Provider disable was introduced for https://bugzilla.redhat.com/show_bug.cgi?id=1437549 to solve another problem. 
When you disable a provider, the workers associated with that provider are not permitted to run. This includes metrics capture, inventory refresh, events and other jobs grabbed by workers. So the provider exists in the DB but the system does not do work for it.
We can use this feature to add many 'work heavy' providers to a running instance but only enable the providers we would like to collect data for at a certain time.

Currently, the system only disables providers before deleting them but we can expose this feature to customers through the UI.

We have to consider that a customer might disable a provider to ease the work load on their system but in the time the provider is disabled we do not collect data for that provider. This might result in data gaps such as missing metrics or events that should have been collected while the provider was disabled.
Comment 2 Ari Zellner 2017-07-30 07:54:43 EDT
Upstream PR: https://github.com/ManageIQ/manageiq-ui-classic/pull/1663
Comment 3 Ari Zellner 2017-08-17 11:32:59 EDT
Followup PR: https://github.com/ManageIQ/manageiq-ui-classic/pull/1950
Comment 5 Pavel Zagalsky 2017-10-30 11:44:19 EDT
After pausing the Containers Provider I see a couple of Collectors still running.
Here are their details:
Name: C&U Metrics Collector for OpenShift
URI: openshift
They still demand some memory usage and CPU.
Is this a normal behaviour?
Comment 6 Jaroslav Henner 2017-11-28 09:23:12 EST
I just have some questions here because I am not sure what is the target here.
 * Which system is loaded by the metrcis collecion? CFME or the provider?
 * Is it not better to just work on getting the collection more lightweight?
 * What is the expected behaviour when there are missing data points? How it is displayed? How it affects averages, maxs, counters?

In general I am quite suspicious about this feature as you can see.
Comment 7 Jaroslav Henner 2017-11-28 09:26:49 EST
I also do not like the nomenclature here. "Disabled provider" means that it is not working. What about "disconnected"?
Comment 8 Jaroslav Henner 2017-11-28 09:37:57 EST
(In reply to Jaroslav Henner from comment #7)
> I also do not like the nomenclature here. "Disabled provider" means that it
> is not working. What about "disconnected"?

I've read part of the discussion about Github PR 1664 about the nomenclature. The image is not enough. When we were talking about this feature on meeting I was confused what Pavel Zagalsky was testing -- disabling a provider? What? How? 

"Disconnected" describes it much better which is important when two colleagues are talking to each other. No one draws icons when talking to other people.
Comment 9 Einat Pacifici 2017-11-30 05:39:35 EST
Moving this back to Verified. Jaroslav, if needed, please open a new BZ with your comments. This specific RFE has been tested and verified.

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