Bug 1946232
Summary: | Openshift console can consume very high CPU | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | OpenShift Container Platform | Reporter: | mchebbi <mchebbi> | ||||||
Component: | Console Kubevirt Plugin | Assignee: | Gilad Lekner <glekner> | ||||||
Status: | CLOSED NEXTRELEASE | QA Contact: | Guohua Ouyang <gouyang> | ||||||
Severity: | high | Docs Contact: | |||||||
Priority: | high | ||||||||
Version: | 4.6 | CC: | aos-bugs, dholler, glekner, gouyang, jjerezro, pducai, ppostler, rsdeor, spadgett, tnisan, ycui, yzamir | ||||||
Target Milestone: | --- | ||||||||
Target Release: | 4.11.0 | ||||||||
Hardware: | Unspecified | ||||||||
OS: | Unspecified | ||||||||
Whiteboard: | Scrubbed | ||||||||
Fixed In Version: | Doc Type: | If docs needed, set a value | |||||||
Doc Text: | Story Points: | --- | |||||||
Clone Of: | Environment: | ||||||||
Last Closed: | 2022-05-16 06:11:01 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: | |||||||||
Attachments: |
|
Description
mchebbi@redhat.com
2021-04-05 13:17:55 UTC
Started the review process but dint have time to finish due to work on stories and bugs. Hey Sam, @mchebbi while investigating the issue, I created 600 VMs, and 600 PVCs to measure the performance between the views. I did notice a lower performance on the VMs page while making a search query. this could happen for couple of reasons: 1. The table has more columns which are more complex to render 2. The VMs page has another tab for VM Templates which has a table of it self, could cause a more complex rendering aswell. 3. All the rows are rendered, IMO we should introduce infinite scrolling or pagination in the future, which will massively improve performance thoughts? (In reply to Gilad Lekner from comment #3) > Hey Sam, @mchebbi > while investigating the issue, I created 600 VMs, and 600 PVCs to measure > the performance between the views. > I did notice a lower performance on the VMs page while making a search query. > this could happen for couple of reasons: > > 1. The table has more columns which are more complex to render > 2. The VMs page has another tab for VM Templates which has a table of it > self, could cause a more complex rendering aswell. > 3. All the rows are rendered, IMO we should introduce infinite scrolling or > pagination in the future, which will massively improve performance > > thoughts? @Gilad Lekner: I totally agree with you, infinite scrolling or pagination could improve performance. Created attachment 1788642 [details]
load virtualization tab over 600 VMs
Create 600 VMs(status is off) on OCP 4.8, the virtualization tab is able to load and the PC which used to launch the web console is not in high CPU.
@mchebbi hey, https://github.com/openshift/console/pull/8821 is merged. can you please check to see if the issue is fixed? In comment #5, we could not reproduce the issue, we will move to next release to debug in the future. pr is open, commenting for redundant needinfo Hi, Is there any chance to ask the customer to verify this bug? I see good results on 4.9.0-fc.0 by loading 134 running VMIs, less than 1s. Since I could not reproduce the bug, I'm not sure whether my tests are enough. Thanks, Guohua moving back to assigned for an additional optimization Created attachment 1839110 [details]
navigate with 385 running VMs
Tested in a cluster which has 385 running VMs and 200+ off VMs, navigate in the virtualization and template page has no issues.
Move the bug to verified based on the test results, feel free to reopen it when later bugs is found.
Could still see the issue on the cluster with the latest console(build from master): https://console-openshift-console.apps.ocp-virt.prod.psi.redhat.com/dashboards. 1. CPU consumer is high when viewing dashboards. 2. CUP consumer is high when viewing VM list page. 3. CPU consumer is also high when viewing Workloads -> Pods. Based on 1 and 3, I guess this is a common OCP or web interface issue because it requests too many resources from the backend at once. Move the bug back to assigned for further investigation. @gouyang @yzamir Created another PR https://github.com/openshift/console/pull/10572 In Attempt to further optimize the VM list status rendering. The Lazy VM Status won't be computed (before it was lazy computed) if the printable status is one of the following: printableStatus === VMStatusSimpleLabel.Paused || printableStatus === VMStatusSimpleLabel.Unknown || printableStatus === VMStatusSimpleLabel.Stopping || printableStatus === VMStatusSimpleLabel.Stopped || printableStatus === VMStatusSimpleLabel.Running Customer closed the customer support, https://access.redhat.com/support/cases/#/case/02902867/discussion?commentId=a0a2K00000ck2y9QAA I prefere not to backport more fixes to 4.10, Guohua, hi can we verify this bug based on customer closing the support case as resolved ? I agree we close the bug as customer close the case already. Set the resolution to NextRelease as it has good results on 4.10 The needinfo request[s] on this closed bug have been removed as they have been unresolved for 365 days |