Add verbose logs for node/plugin scores even ranged in low levels #103515

muma378 · 2021-07-06T12:54:13Z

What type of PR is this?

/kind feature

What this PR does / why we need it:

No scheduling info captured if the logging level less than 10. It would be a little help to find what happend if the top N highest node/plugin scores dumped at a lower level.

@damemi made some improvement in #99411 , however it is found ~50% performace drop in large cluster.

As @Huang-Wei suggests in #101820，we use a dynamical logging strategy to dump depending on the logging level:

for logging level ranged in [4, 6), dump the scores for topM nodes/plugins. And make it as performant as possible.
for logging level ranged in [6, 10), dump the scores for topN nodes, and probably show all plugins
for logging level >= 10, dump the scores for all nodes/plugins.

The output looks like this with -v=4:

I0728 15:09:41.627395   98057 scheduling_queue.go:872] "About to try and schedule pod" pod="scheduler-extenderbd2630b2-df6f-4a45-abe0-7e80e26ac76d/extender-test-pod"
I0728 15:09:41.627418   98057 scheduler.go:516] "Attempting to schedule pod" pod="scheduler-extenderbd2630b2-df6f-4a45-abe0-7e80e26ac76d/extender-test-pod"
E0728 15:09:41.627519   98057 plugin.go:138] "getting namespace, assuming empty set of namespace labels" err="namespace \"scheduler-extenderbd2630b2-df6f-4a45-abe0-7e80e26ac76d\" not found" namespace="scheduler-extenderbd2630b2-df6f-4a45-
abe0-7e80e26ac76d"
E0728 15:09:41.630656   98057 plugin.go:138] "getting namespace, assuming empty set of namespace labels" err="namespace \"scheduler-extenderbd2630b2-df6f-4a45-abe0-7e80e26ac76d\" not found" namespace="scheduler-extenderbd2630b2-df6f-4a45-
abe0-7e80e26ac76d"
I0728 15:09:41.630850   98057 generic_scheduler.go:597] "Top 3 plugins for pod on node" pod="scheduler-extenderbd2630b2-df6f-4a45-abe0-7e80e26ac76d/extender-test-pod" node="machine3" score=1000400 plugins=[{Name:NodePreferAvoidPods Score:
1000000} {Name:PodTopologySpread Score:200} {Name:TaintToleration Score:100}]
I0728 15:09:41.630875   98057 generic_scheduler.go:597] "Top 3 plugins for pod on node" pod="scheduler-extenderbd2630b2-df6f-4a45-abe0-7e80e26ac76d/extender-test-pod" node="machine2" score=1000400 plugins=[{Name:NodePreferAvoidPods Score:
1000000} {Name:PodTopologySpread Score:200} {Name:TaintToleration Score:100}]

The output looks like this with -v=6:

I0728 15:01:01.121359   96269 scheduler.go:516] "Attempting to schedule pod" pod="scheduler-extender87c5ecb8-b5bf-4723-b247-3905865f6a5d/extender-test-pod"
E0728 15:01:01.121444   96269 plugin.go:138] "getting namespace, assuming empty set of namespace labels" err="namespace \"scheduler-extender87c5ecb8-b5bf-4723-b247-3905865f6a5d\" not found" namespace="scheduler-extender87c5ecb8-b5bf-4723-
b247-3905865f6a5d"
E0728 15:01:01.123943   96269 plugin.go:138] "getting namespace, assuming empty set of namespace labels" err="namespace \"scheduler-extender87c5ecb8-b5bf-4723-b247-3905865f6a5d\" not found" namespace="scheduler-extender87c5ecb8-b5bf-4723-
b247-3905865f6a5d"
I0728 15:01:01.124131   96269 generic_scheduler.go:597] "Top 10 plugins for pod on node" pod="scheduler-extender87c5ecb8-b5bf-4723-b247-3905865f6a5d/extender-test-pod" node="machine2" score=1000400 plugins=[{Name:NodePreferAvoidPods Score
:1000000} {Name:PodTopologySpread Score:200} {Name:NodeResourcesBalancedAllocation Score:100} {Name:TaintToleration Score:100} {Name:InterPodAffinity Score:0} {Name:NodeResourcesLeastAllocated Score:0} {Name:NodeAffinity Score:0} {Name:Im
ageLocality Score:0}]
I0728 15:01:01.124156   96269 generic_scheduler.go:597] "Top 10 plugins for pod on node" pod="scheduler-extender87c5ecb8-b5bf-4723-b247-3905865f6a5d/extender-test-pod" node="machine3" score=1000400 plugins=[{Name:NodePreferAvoidPods Score
:1000000} {Name:PodTopologySpread Score:200} {Name:NodeResourcesBalancedAllocation Score:100} {Name:TaintToleration Score:100} {Name:InterPodAffinity Score:0} {Name:NodeResourcesLeastAllocated Score:0} {Name:NodeAffinity Score:0} {Name:Im
ageLocality Score:0}]

Which issue(s) this PR fixes:

Fixes #101820

Special notes for your reviewer:

Performance was evaluated and compared with the tool k8s-sched-perf-stat:

I did the performance test with the following 4 scenarios（2 branches x 2 logging levels）:

	v=1	v=4
master	perf-v1-clean.txt	perf-v4-clean.txt
feature	perf-v1-verbose.txt	perf-v4-verbose.txt

with the command:

make test-integration WHAT=./test/integration/scheduler_perf KUBE_TEST_VMODULE=\"$KUBE_TEST_VMODULE\" KUBE_TEST_ARGS="-alsologtostderr=false -logtostderr=false -benchtime=1ns -bench=BenchmarkPerfScheduling/SchedulingBasic/5000Nodes/5000InitPods/1000PodsToSchedule -data-items-dir /tmp"

and the diff is evelauted:

++ bash -c 'go run *.go /Users/xiaoyang/go/src/k8s.io/kubernetes/perf-v1-clean.txt /Users/xiaoyang/go/src/k8s.io/kubernetes/perf-v1-verbose.txt'
+--------------------------+---------------------------+----------+--------+---------+---------+----------+
|          METRIC          |           GROUP           | QUANTILE |  UNIT  |   OLD   |   NEW   |   DIFF   |
+--------------------------+---------------------------+----------+--------+---------+---------+----------+
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Average  | pods/s |  106.19 |  130.95 | +23.32%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc50   | pods/s |   92.11 |   98.62 | +7.07%   |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc90   | pods/s |  217.78 |  275.38 | +26.45%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc99   | pods/s |  229.00 |  275.38 | +20.25%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Average  | ms     | 1202.25 | 1237.40 | +2.92%   |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc50   | ms     |  348.70 |  786.11 | +125.44% |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc90   | ms     | 3358.05 | 2956.64 | -11.95%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc99   | ms     | 7474.71 | 6202.31 | -17.02%  |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Average  | ms     | 1360.00 | 1348.09 | -0.88%   |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc50   | ms     |  607.56 |  981.18 | +61.50%  |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc90   | ms     | 4041.90 | 3599.99 | -10.93%  |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc99   | ms     | 8367.33 | 6907.92 | -17.44%  |
+--------------------------+---------------------------+----------+--------+---------+---------+----------+
++ bash -c 'go run *.go /Users/xiaoyang/go/src/k8s.io/kubernetes/perf-v4-clean.txt /Users/xiaoyang/go/src/k8s.io/kubernetes/perf-v4-verbose.txt'
+--------------------------+---------------------------+----------+--------+----------+---------+----------+
|          METRIC          |           GROUP           | QUANTILE |  UNIT  |   OLD    |   NEW   |   DIFF   |
+--------------------------+---------------------------+----------+--------+----------+---------+----------+
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Average  | pods/s |   105.88 |  138.51 | +30.82%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc50   | pods/s |    80.25 |   97.62 | +21.65%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc90   | pods/s |   223.13 |  275.67 | +23.55%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc99   | pods/s |   234.75 |  277.22 | +18.09%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Average  | ms     |   715.89 |  865.39 | +20.88%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc50   | ms     |    47.56 |  386.74 | +713.23% |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc90   | ms     |  2282.13 | 2510.42 | +10.00%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc99   | ms     |  5320.34 | 5314.16 | -0.12%   |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Average  | ms     |  4508.74 | 1922.55 | -57.36%  |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc50   | ms     |  4688.92 | 1747.25 | -62.74%  |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc90   | ms     |  8867.54 | 3840.48 | -56.69%  |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc99   | ms     | 11172.94 | 6716.56 | -39.89%  |
+--------------------------+---------------------------+----------+--------+----------+---------+----------+
++ bash -c 'go run *.go /Users/xiaoyang/go/src/k8s.io/kubernetes/perf-v1-clean.txt /Users/xiaoyang/go/src/k8s.io/kubernetes/perf-v4-clean.txt'
+--------------------------+---------------------------+----------+--------+---------+----------+----------+
|          METRIC          |           GROUP           | QUANTILE |  UNIT  |   OLD   |   NEW    |   DIFF   |
+--------------------------+---------------------------+----------+--------+---------+----------+----------+
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Average  | pods/s |  106.19 |   105.88 | -0.29%   |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc50   | pods/s |   92.11 |    80.25 | -12.88%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc90   | pods/s |  217.78 |   223.13 | +2.46%   |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc99   | pods/s |  229.00 |   234.75 | +2.51%   |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Average  | ms     | 1202.25 |   715.89 | -40.45%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc50   | ms     |  348.70 |    47.56 | -86.36%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc90   | ms     | 3358.05 |  2282.13 | -32.04%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc99   | ms     | 7474.71 |  5320.34 | -28.82%  |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Average  | ms     | 1360.00 |  4508.74 | +231.53% |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc50   | ms     |  607.56 |  4688.92 | +671.77% |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc90   | ms     | 4041.90 |  8867.54 | +119.39% |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc99   | ms     | 8367.33 | 11172.94 | +33.53%  |
+--------------------------+---------------------------+----------+--------+---------+----------+----------+
++ bash -c 'go run *.go /Users/xiaoyang/go/src/k8s.io/kubernetes/perf-v1-verbose.txt /Users/xiaoyang/go/src/k8s.io/kubernetes/perf-v4-verbose.txt'
+--------------------------+---------------------------+----------+--------+---------+---------+---------+
|          METRIC          |           GROUP           | QUANTILE |  UNIT  |   OLD   |   NEW   |  DIFF   |
+--------------------------+---------------------------+----------+--------+---------+---------+---------+
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Average  | pods/s |  130.95 |  138.51 | +5.77%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc50   | pods/s |   98.62 |   97.62 | -1.01%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc90   | pods/s |  275.38 |  275.67 | +0.11%  |
| SchedulingThroughput     | SchedulingBasic/5000Nodes | Perc99   | pods/s |  275.38 |  277.22 | +0.67%  |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Average  | ms     | 1237.40 |  865.39 | -30.06% |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc50   | ms     |  786.11 |  386.74 | -50.80% |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc90   | ms     | 2956.64 | 2510.42 | -15.09% |
| scheduler_e2e_scheduling | SchedulingBasic/5000Nodes | Perc99   | ms     | 6202.31 | 5314.16 | -14.32% |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Average  | ms     | 1348.09 | 1922.55 | +42.61% |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc50   | ms     |  981.18 | 1747.25 | +78.08% |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc90   | ms     | 3599.99 | 3840.48 | +6.68%  |
| scheduler_pod_scheduling | SchedulingBasic/5000Nodes | Perc99   | ms     | 6907.92 | 6716.56 | -2.77%  |
+--------------------------+---------------------------+----------+--------+---------+---------+---------+

Does this PR introduce a user-facing change?

kube-scheduler now logs node and plugin scoring  even though --v<10
- socres of the top 3 plugins in the top 3 nodes are dumped if --v=4,5
- socres of all plugins in the top 6 nodes are dumped if --v=6,7,8,9

Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:

k8s-ci-robot · 2021-07-06T12:54:20Z

@muma378: This issue is currently awaiting triage.

If a SIG or subproject determines this is a relevant issue, they will accept it by applying the triage/accepted label and provide further guidance.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

k8s-ci-robot · 2021-07-06T12:54:21Z

Hi @muma378. Thanks for your PR.

I'm waiting for a kubernetes member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

wzshiming · 2021-07-07T01:43:06Z

/ok-to-test

muma378 · 2021-07-07T03:08:07Z

/retest

damemi · 2021-07-07T13:23:34Z

@muma378 is this ready for review? If so please move it from "draft" to ready so we know it's not still a WIP :)

muma378 · 2021-07-09T17:50:51Z

@muma378 is this ready for review? If so please move it from "draft" to ready so we know it's not still a WIP :)

@damemi very thanks for your remind, but i‘m still strugging in running e2e-test in my local laptop to check if the logs were output correctly.
Do you have any idea to help me to get the kube-scheduler logs without building a new cluster? any suggest would be appreciate.

pkg/scheduler/generic_scheduler.go

muma378 · 2021-09-01T01:52:50Z

@damemi @ahg-g PTAL

damemi

Sorry for the delay on this @muma378. This looks pretty good to me, do I understand the performance results correctly that this improved the performance at all log levels? That's really interesting, is it due to the use of the heap do you think?

I think this is good to go, but will let @Huang-Wei do final lgtm since he caught the issue last time
/approve

k8s-ci-robot · 2021-09-07T14:44:46Z

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: damemi, muma378

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

~~pkg/scheduler/OWNERS~~ [damemi]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

wzshiming · 2021-09-08T03:14:47Z

/lgtm

Huang-Wei · 2021-09-08T05:11:04Z

/lgtm

Well, I think this made this PR merged without thorough reviewing. @wzshiming please hold off the directive /lgtm if it needs further discussion.

After taking a glance, I can spot several issues:

A consensus on the value M and N. This is not fully discussed yet.

Further dig into the performance result. The test runs on v=1 and v=4, but actually they're the same if it's running in the context of integration framework.

kubernetes/test/integration/framework/master_utils.go

Lines 131 to 136 in 90851a0

    
           // Ensure we log at least level 4 
        
           v := flag.Lookup("v").Value 
        
           level, _ := strconv.Atoi(v.String()) 
        
           if level < 4 { 
        
           	v.Set("4") 
        
           }

I'd expect to see the result comparing before/after on v=4 and v=6.

Further review on the code, like adding UTs and some logic nits.

wzshiming · 2021-09-08T05:57:32Z

@Huang-Wei
I am sorry, I didn't notice that the review was /approve, and /lgtm was added This only represents me personally.

pacoxu · 2021-09-08T06:10:47Z

I suggest reopening the issue and follow up on Wei's comments.

/cc @kerthcet @muma378

k8s-ci-robot · 2021-09-08T06:10:48Z

@pacoxu: GitHub didn't allow me to request PR reviews from the following users: muma378.

Note that only kubernetes members and repo collaborators can review this PR, and authors cannot review their own PRs.

In response to this:

I suggest reopening the issue and follow up on Wei's comments.

/cc @kerthcet @muma378

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

Huang-Wei · 2021-09-08T16:15:54Z

@damemi @ahg-g we may have to revert it and open a new PR for a thorough review, WDYT?

damemi · 2021-09-08T16:17:26Z

I didn't realize the performance tests didn't cover all the cases, sorry. We should try to at least get some measurements for that first. If there is still a big hit to performance then we should revert

ahg-g · 2021-09-08T17:15:10Z

+1 to reverting. The logic seems a bit ad hoc to me and not necessarily placed in the right pkg.

If we are serious about dumping internal scheduler state or decisions, we need to design a library for that, which includes the logic we currently have for dumping the cache:

Huang-Wei · 2021-09-08T20:31:11Z

@damemi I created #104849 to revert this PR.

@muma378 Apologize for the reverting. It doesn't we don't accept the changes. It's just we want to review it thoroughly. Feel free to open a new PR, taking the comments #103515 (comment) and #103515 (comment) into consideration.

muma378 · 2021-09-09T05:10:36Z

@damemi I created #104849 to revert this PR.

@muma378 Apologize for the reverting. It doesn't we don't accept the changes. It's just we want to review it thoroughly. Feel free to open a new PR, taking the comments #103515 (comment) and #103515 (comment) into consideration.

@Huang-Wei Never mind, it is absolutely ok with me. I feel sorry about the mistake on performance testing and will keep following the issue.

In terms of your and @ahg-g 's suggestion:

I will do more performance test with logging levels 5/6/9, and add UTs;
Is there anything I can do about the value choosing of M and N? (such as comparing the performace drop between different values?)
Add verbose logs for node/plugin scores even ranged in low levels #103515 (comment) sounds reasonable, however it may be a big work. Should I just open a new PR or start it in a more formal way (such as writting a design proposal)?

Add verbose logs for node/plugin scores when scheduling

1e2bf25

k8s-ci-robot added needs-triage Indicates an issue or PR lacks a `triage/foo` label and requires one. needs-priority Indicates a PR lacks a `priority/foo` label and requires one. labels Jul 6, 2021

k8s-ci-robot requested review from ahg-g and damemi July 6, 2021 12:55

k8s-ci-robot added ok-to-test Indicates a non-member PR verified by an org member that is safe to test. and removed needs-ok-to-test Indicates a PR that requires an org member to verify it is safe to test. labels Jul 7, 2021

yuzhiquan reviewed Jul 27, 2021

View reviewed changes

pkg/scheduler/generic_scheduler.go Outdated Show resolved Hide resolved

yuzhiquan reviewed Jul 27, 2021

View reviewed changes

pkg/scheduler/generic_scheduler.go Outdated Show resolved Hide resolved

kerthcet reviewed Jul 27, 2021

View reviewed changes

pkg/scheduler/generic_scheduler.go Show resolved Hide resolved

Refactor the structure nodeScoreHeap

144f25a

k8s-ci-robot added release-note Denotes a PR that will be considered when it comes time to generate release notes. and removed release-note-none Denotes a PR that doesn't merit a release note. labels Jul 28, 2021

muma378 marked this pull request as ready for review July 28, 2021 07:15

k8s-ci-robot removed the do-not-merge/work-in-progress Indicates that a PR should not merge because it is a work in progress. label Jul 28, 2021

muma378 requested a review from yuzhiquan July 28, 2021 07:17

kerthcet reviewed Jul 28, 2021

View reviewed changes

pkg/scheduler/generic_scheduler.go Show resolved Hide resolved

damemi reviewed Sep 7, 2021

View reviewed changes

k8s-ci-robot added the approved Indicates a PR has been approved by an approver from all required OWNERS files. label Sep 7, 2021

k8s-ci-robot assigned wzshiming Sep 8, 2021

k8s-ci-robot added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Sep 8, 2021

k8s-ci-robot merged commit 3282d6c into kubernetes:master Sep 8, 2021

k8s-ci-robot added this to the v1.23 milestone Sep 8, 2021

k8s-ci-robot requested a review from kerthcet September 8, 2021 06:10

Huang-Wei mentioned this pull request Sep 8, 2021

Revert PR 103515 #104849

Merged

pacoxu mentioned this pull request Sep 9, 2021

More debug logs to check scheduler spread algorithm #101820

Closed

sanposhiho mentioned this pull request Jan 24, 2022

Update CHANGELOG-1.23.md to delete reverted change #107712

Merged

Huang-Wei mentioned this pull request Sep 8, 2022

add(scheduler): add the benchmark test for the schedulePod function #112220

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add verbose logs for node/plugin scores even ranged in low levels #103515

Add verbose logs for node/plugin scores even ranged in low levels #103515

muma378 commented Jul 6, 2021 •

edited

k8s-ci-robot commented Jul 6, 2021

k8s-ci-robot commented Jul 6, 2021

wzshiming commented Jul 7, 2021

muma378 commented Jul 7, 2021

damemi commented Jul 7, 2021

muma378 commented Jul 9, 2021

muma378 commented Sep 1, 2021

damemi left a comment

k8s-ci-robot commented Sep 7, 2021

wzshiming commented Sep 8, 2021

Huang-Wei commented Sep 8, 2021

wzshiming commented Sep 8, 2021 •

edited

pacoxu commented Sep 8, 2021

k8s-ci-robot commented Sep 8, 2021

Huang-Wei commented Sep 8, 2021

damemi commented Sep 8, 2021

ahg-g commented Sep 8, 2021

Huang-Wei commented Sep 8, 2021

muma378 commented Sep 9, 2021

Add verbose logs for node/plugin scores even ranged in low levels #103515

Add verbose logs for node/plugin scores even ranged in low levels #103515

Conversation

muma378 commented Jul 6, 2021 • edited

What type of PR is this?

What this PR does / why we need it:

Which issue(s) this PR fixes:

Special notes for your reviewer:

Does this PR introduce a user-facing change?

Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:

k8s-ci-robot commented Jul 6, 2021

k8s-ci-robot commented Jul 6, 2021

wzshiming commented Jul 7, 2021

muma378 commented Jul 7, 2021

damemi commented Jul 7, 2021

muma378 commented Jul 9, 2021

muma378 commented Sep 1, 2021

damemi left a comment

Choose a reason for hiding this comment

k8s-ci-robot commented Sep 7, 2021

wzshiming commented Sep 8, 2021

Huang-Wei commented Sep 8, 2021

wzshiming commented Sep 8, 2021 • edited

pacoxu commented Sep 8, 2021

k8s-ci-robot commented Sep 8, 2021

Huang-Wei commented Sep 8, 2021

damemi commented Sep 8, 2021

ahg-g commented Sep 8, 2021

Huang-Wei commented Sep 8, 2021

muma378 commented Sep 9, 2021

muma378 commented Jul 6, 2021 •

edited

wzshiming commented Sep 8, 2021 •

edited