Human readable memory for VPA

[aleangelin88]

Hello, I have 2 questions that I cannot find answers for online

question 1: is it possible to have the indication of memory in a more human readable fashion? like in Kb, Mb, Gb, etc

NAMESPACE NAME MODE CPU MEM PROVIDED AGE default analyzer-vpa Off 1038m 4992073454 True 72m default input-manager-vpa Off 25m 262144k True 72m

question 2: is the indication of the cpu and memory usage per pods or per deployment? if I activate the autoscaler, would it divide the request and limit per pods accordingly?

thank you in advance

[adrianmoisey]

/area vertical-pod-autoscaler

[austinorth]

Would also love human-readable memory!

[adrianmoisey]

/retitle Human readable memory for VPA

[adrianmoisey]

This has been merged to master, and should be available in the next VPA release. /close

Thanks @omerap12 for doing this one!

[k8s-ci-robot]

@adrianmoisey: Closing this issue.

In response to this:

This has been merged to master, and should be available in the next VPA release. /close

Thanks @omerap12 for doing this one!

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-sigs/prow repository.

[omerap12]

This has been merged to master, and should be available in the next VPA release.

/close

Thanks @omerap12 for doing this one!

My pleasure 😀

[kyleli666]

Hi all, I tried with 1.3.0 image with "--humanize-memory=true" today, but found the MEM value now having more digits with the unit "milibyte". This value looks to align with the code design 272063528960/1000/1024/1024=259.46Mi, but seems more unreadable. Is there some configuration I missed?

$ k get verticalpodautoscalers.autoscaling.k8s.io -n logging fluent-bit-service
NAME                 MODE      CPU    MEM             PROVIDED   AGE
fluent-bit-service   Initial   126m   272063528960m   True       2d21h

[adrianmoisey]

Hi all, I tried with 1.3.0 image with "--humanize-memory=true" today, but found the MEM value now having more digits with the unit "milibyte". This value looks to align with the code design 272063528960/1000/1024/1024=259.46Mi, but seems more unreadable. Is there some configuration I missed?

Can you give us some steps to reproduce this? I tested it and I got the correct result:

$ k get vpa -w
NAME          MODE   CPU   MEM   PROVIDED   AGE
hamster-vpa   Auto                          8s
hamster-vpa   Auto   627m   250Mi   True       48s
hamster-vpa   Auto   627m   250Mi   True       108s

I then turned the setting off, and the recommendation reverted to the non-human-readable format:

$ k get vpa -w
NAME          MODE   CPU    MEM     PROVIDED   AGE
hamster-vpa   Auto   627m   250Mi   True       2m22s
hamster-vpa   Auto   627m   262144k   True       3m23s
hamster-vpa   Auto   627m   262144k   True       4m23s

[kyleli666]

Hi @adrianmoisey , my vpa is installed with

tag=1.3.0; helm upgrade -i -n kube-system vpa fairwinds-stable/vpa -f vpa-values.yaml --set recommender.image.tag=$tag,updater.image.tag=$tag,admissionController.image.tag=$tag

# vpa-values.yaml
recommender:
  # recommender.enabled -- If true, the vpa recommender component will be installed.
  enabled: true
  # recommender.extraArgs -- A set of key-value flags to be passed to the recommender
  extraArgs:
    v: "4"
    humanize-memory: true # starting from 1.3.0
    pod-recommendation-min-cpu-millicores: 10
    pod-recommendation-min-memory-mb: 50
    target-cpu-percentile: 0.50
    target-memory-percentile: 0.50
  replicaCount: 1

Without "humanize-memory: true", I get this result, seems already worse than yours :\

$ k get vpa -A -w
NAMESPACE     NAME                       MODE      CPU    MEM         PROVIDED   AGE
kube-system   vpa-admission-controller   Auto      11m    52428800    True       158d
kube-system   vpa-recommender            Auto      11m    63544758    True       158d
kube-system   vpa-updater                Auto      11m    78221997    True       158d
logging       fluent-bit-service         Initial   126m   225384266   True       3d2h
monitoring    grafana                    Auto      11m    93633096    True       146d

With "humanize-memory: true", I get a result where 50Mi looks good but the others not.

$ k get vpa -A -w
NAMESPACE     NAME                       MODE      CPU    MEM             PROVIDED   AGE
kube-system   vpa-admission-controller   Auto      11m    50Mi            True       158d
kube-system   vpa-recommender            Auto      11m    63543705600m    True       158d
kube-system   vpa-updater                Auto      11m    93637836800m    True       158d
logging       fluent-bit-service         Initial   126m   272063528960m   True       3d2h
monitoring    grafana                    Auto      11m    93637836800m    True       146d

Thanks for your help!

[omerap12]

Hi @adrianmoisey , my vpa is installed with

tag=1.3.0; helm upgrade -i -n kube-system vpa fairwinds-stable/vpa -f vpa-values.yaml --set recommender.image.tag=$tag,updater.image.tag=$tag,admissionController.image.tag=$tag

# vpa-values.yaml
recommender:
  # recommender.enabled -- If true, the vpa recommender component will be installed.
  enabled: true
  # recommender.extraArgs -- A set of key-value flags to be passed to the recommender
  extraArgs:
    v: "4"
    humanize-memory: true # starting from 1.3.0
    pod-recommendation-min-cpu-millicores: 10
    pod-recommendation-min-memory-mb: 50
    target-cpu-percentile: 0.50
    target-memory-percentile: 0.50
  replicaCount: 1

Without "humanize-memory: true", I get this result, seems already worse than yours :\

$ k get vpa -A -w
NAMESPACE     NAME                       MODE      CPU    MEM         PROVIDED   AGE
kube-system   vpa-admission-controller   Auto      11m    52428800    True       158d
kube-system   vpa-recommender            Auto      11m    63544758    True       158d
kube-system   vpa-updater                Auto      11m    78221997    True       158d
logging       fluent-bit-service         Initial   126m   225384266   True       3d2h
monitoring    grafana                    Auto      11m    93633096    True       146d

With "humanize-memory: true", I get a result where 50Mi looks good but the others not.

$ k get vpa -A -w
NAMESPACE     NAME                       MODE      CPU    MEM             PROVIDED   AGE
kube-system   vpa-admission-controller   Auto      11m    50Mi            True       158d
kube-system   vpa-recommender            Auto      11m    63543705600m    True       158d
kube-system   vpa-updater                Auto      11m    93637836800m    True       158d
logging       fluent-bit-service         Initial   126m   272063528960m   True       3d2h
monitoring    grafana                    Auto      11m    93637836800m    True       146d

Thanks for your help!

Can you share the recommender logs? We print wether we convert the metrics or not

[adrianmoisey]

Can you paste your VPA objects here too please? At least for the VPA components.

[adrianmoisey]

Ah, I reproduced it by increasing memory usage:

$ k get vpa -w -A
NAMESPACE     NAME              MODE   CPU    MEM    PROVIDED   AGE
default       hamster-vpa       Auto   100m   50Mi   True       3m2s
kube-system   vpa-recommender   Auto   10m    50Mi   True       110s
default       hamster-vpa       Auto   100m   183049912320m   True       4m42s

I didn't get a log similar to:

I0126 08:28:31.646220       1 types.go:99] "Converting raw value to humanized value" rawValue=52428800 humanizedValue="50.00Mi"

[kyleli666]

@omerap12 my logs recommender.log

@adrianmoisey My VPA Objects for the VPA components.

---
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: vpa-admission-controller
  namespace: kube-system
spec:
  targetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: vpa-admission-controller
  updatePolicy:
    updateMode: Auto
---
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: vpa-recommender
  namespace: kube-system
spec:
  targetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: vpa-recommender
  updatePolicy:
    updateMode: Auto
  resourcePolicy:
    containerPolicies:
      - containerName: '*'
        minAllowed:
          memory: 50Mi
---
apiVersion: autoscaling.k8s.io/v1
kind: VerticalPodAutoscaler
metadata:
  name: vpa-updater
  namespace: kube-system
spec:
  targetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: vpa-updater
  updatePolicy:
    updateMode: Auto

[adrianmoisey]

@kyleli666 could you open a new bug issue for this? It's strange to keep discussion on a closed issue

[kyleli666]

@kyleli666 could you open a new bug issue for this? It's strange to keep discussion on a closed issue

Here it is https://github.com/kubernetes/autoscaler/issues/7770