Right Sizing recommendations with VPA¶

Recommendations example¶

← On the left side you can see:

The initial values that were set before Virtual Pod Autoscaler (VPA) took action.
One of the pods is being terminated, due to auto mode changing the CPU/MEM values right away (causing potential downtime).

→ On the right side you can see:

VPA recommendation couple minutes right after.

VPA install guide¶

In order for the instructions to work on macOS one needs to have latest version of OpenSSL from Homebrew:
- brew install openssl (make use of that OpenSSL version, rather than macOS native one).
- Run brew info openssl to see the instructions for setting this openssl version as your default one.
Metrics-server up and running
- If you don't have metrics-server installed, run the following to install it:
```
kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml
```
- Run kubectl top nodes to check if metrics server was installed successfully

Vertical Pod Autoscaler installed

Prerequisite: helm
Install helm chart

helm repo add fairwinds-stable https://charts.fairwinds.com/stable
helm install vpa fairwinds-stable/vpa --namespace vpa --create-namespace

Run helm chart tests to make sure VPA is successfully installed

helm test vpa -n vpa

Output similar to this should be visible:

LAST DEPLOYED: Thu Apr 29 19:53:50 2021
NAMESPACE: vpa
STATUS: deployed
REVISION: 1
TEST SUITE:     vpa-test
Last Started:   Thu Apr 29 20:04:13 2021
Last Completed: Thu Apr 29 20:04:13 2021
Phase:          Succeeded
TEST SUITE:     vpa-test
Last Started:   Thu Apr 29 20:04:14 2021
Last Completed: Thu Apr 29 20:04:14 2021
Phase:          Succeeded
TEST SUITE:     vpa-test
Last Started:   Thu Apr 29 20:04:14 2021
Last Completed: Thu Apr 29 20:04:14 2021
Phase:          Succeeded
TEST SUITE:     vpa-checkpoint-crd-available
Last Started:   Thu Apr 29 20:04:14 2021
Last Completed: Thu Apr 29 20:04:16 2021
Phase:          Succeeded
TEST SUITE:     vpa-crd-available
Last Started:   Thu Apr 29 20:04:16 2021
Last Completed: Thu Apr 29 20:04:19 2021
Phase:          Succeeded
TEST SUITE:     vpa-test-create-vpa
Last Started:   Thu Apr 29 20:04:21 2021
Last Completed: Thu Apr 29 20:04:49 2021
Phase:          Succeeded
TEST SUITE:     vpa-metrics-api-available
Last Started:   Thu Apr 29 20:04:19 2021
Last Completed: Thu Apr 29 20:04:21 2021
Phase:          Succeeded
NOTES:
Congratulations on installing the Vertical Pod Autoscaler!

Components Installed:
  - recommender
  - updater

Create VPA for each deployment to get the recommendations
- See Configure VPA for your deployment below for examples.
- Each deployment that wants make use of VPA, needs to have VPA created for it.
- Wait a day for the VPA.
- Send us the output of:
```
kubectl get vpa -A -o yaml > recommendations.txt
```

Troubleshooting¶

In case something goes wrong, detailed instructions can be found at our github.
Make sure that on macOS the openssl from Homebrew is used

Configure VPA for your deployment¶

In order to see how it works, you can use the example bellow, which would create both example deployment & example VPA configuration for that deployment.

Once the deployment yaml is applied using kubectl view recommendations with the following line:

kubectl describe vpa/hamster-vpa

Example output:

Status:
  Conditions:
    Last Transition Time:  2021-04-27T06:13:54Z
    Status:                True
    Type:                  RecommendationProvided
  Recommendation:
    Container Recommendations:
      Container Name:  hamster
      Lower Bound:
        Cpu:     491m
        Memory:  262144k
      Target:
        Cpu:     587m
        Memory:  262144k
      Uncapped Target:
        Cpu:     587m
        Memory:  262144k
      Upper Bound:
        Cpu:     1
        Memory:  262144k
Events:          <none>

Example deployment with VPA:

# This config creates a deployment with two pods, each requesting 100 millicores
# and trying to utilize slightly above 500 millicores (repeatedly using CPU for
# 0.5s and sleeping 0.5s).
# It also creates a corresponding Vertical Pod Autoscaler that adjusts the
# requests.
# Note that the update mode is left unset, so it defaults to "Auto" mode.
---
apiVersion: "autoscaling.k8s.io/v1"
kind: VerticalPodAutoscaler
metadata:
  name: hamster-vpa
  namespace: default
spec:
  targetRef:
    apiVersion: "apps/v1"
    kind: Deployment
    name: hamster
  updatePolicy:
    # updateMode set to Off
    # - runs in recommendation mode
    # - does not mess with your pod request/limit configurations
    updateMode: "Off"
  resourcePolicy:
    containerPolicies:
      - containerName: '*'
        controlledResources: ["cpu", "memory"]
---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: hamster
  namespace: default
spec:
  selector:
    matchLabels:
      app: hamster
  replicas: 2
  template:
    metadata:
      labels:
        app: hamster
    spec:
      securityContext:
        runAsNonRoot: true
        runAsUser: 65534 # nobody
      containers:
        - name: hamster
          image: k8s.gcr.io/ubuntu-slim:0.1
          resources:
            requests:
              cpu: 100m
              memory: 50Mi
          command: ["/bin/sh"]
          args:
            - "-c"
            - "while true; do timeout 0.5s yes >/dev/null; sleep 0.5s; done"

Available VPA modes¶

"Off": VPA does not automatically change resource requirements of the pods. The recommendations are calculated and can be inspected in the VPA object.
"Auto": VPA assigns resource requests on pod creation as well as updates them on existing pods using the preferred update mechanism. Currently this is equivalent to "Recreate" (see below). Once restart free ("in-place") update of pod requests is available, it may be used as the preferred update mechanism by the "Auto" mode.

Warning

"Auto" feature of VPA is experimental and may cause downtime for your applications.

"Initial": VPA only assigns resource requests on pod creation and never changes them later.

Resources¶

VPA definitive guide