Red Hat Bugzilla – Bug 1030095
Platform Utilization portlet does not work with large systems
Last modified: 2015-09-02 20:02:15 EDT
Description of problem:
The portlet/report "Platform Utilization" that shows cpu/memory/swap utilization across all nodes, fails to work once you have more than a few dosin nodes. At best you get time outs, at worst it hangs the whole UI for minutes until you get a popup with "unknown error".
The results needs to be pre-calculated before the report is done, and the results should be filterable by groups for dash-board use. I propose adding a limit of 10 or 20 servers in the current version to avoid having this problem.
Have a system with 50+ hosts in the inventory. Go to reports and choose "Platform Utilization" or add the "Platform Utilization" portlet to a dashboard.
List of platforms and a graphical utilization image per host.
I think this was fixed in 3.2 with the work for Bug 872731. Setting to ON_QA if the team wants to test this use case, or close as desired.
Moving to VERIFIED.
I tested this BZ using 2 scenarios.
In the first one more then 50 mock platforms were added to the inventory -> the UI displayed the Platform Utilization portlet with no error.
The second scenario used slow agent by disabling TCP connection using firewall (see BZ 872731) -> The UI did not lag nor it showed any error message. After each refresh the ERROR message (...Failed to send command...) is logged in server.log