Add Vertical Pod Autoscaler support (#183)

* feat: add Vertical Pod Autoscaler

* docs: add VPA section

* feat: Add tests for VPA

* feat: add support for k8s < 1.25 for VPA

* fix(vpa): fix typos

* fix: typos

charts/k8s-service/README.md

@@ -42,6 +42,9 @@ The following resources will be deployed with this Helm Chart, depending on whic
 - `Horizontal Pod Autoscaler`: The `Horizontal Pod Autoscaler` automatically scales the number of pods in a replication
                                 controller, deployment, replica set or stateful set based on observed CPU or memory utilization.
                                 Created only if the user sets `horizontalPodAutoscaler.enabled = true`.
+- `Vertical Pod Autoscaler`: The `Vertical Pod Autoscaler` can offer recommendations or change the CPU and memory for both 
+                             requests and limits based on specified configuration.
+                              Created only if the user sets `verticalPodAutoscaler.enabled = true`.
 - `PodDisruptionBudget`: The `PodDisruptionBudget` resource that specifies a disruption budget for the `Pods` managed by
                          the `Deployment`. This manages how many pods can be disrupted by a voluntary disruption (e.g
                          node maintenance). Created if you specify a non-zero value for the `minPodsAvailable` input
@@ -1218,3 +1221,39 @@ You can learn more about using private registries with Kubernetes in [the offici
 documentation](https://kubernetes.io/docs/concepts/containers/images/#using-a-private-registry).
 
 back to [root README](/README.adoc#day-to-day-operations)
+
+## How to enable Vertical Pod Autoscaler ?
+
+[Vertical Pod Auto scaler](https://github.com/kubernetes/autoscaler/tree/master/vertical-pod-autoscaler) is used to dynamically change
+the requests and limits of running pods. It should **not be used** in combination with Horizontal Pod Autoscaler.
+
+First you need to install it in the cluster by following the [installation guide](https://github.com/kubernetes/autoscaler/tree/master/vertical-pod-autoscaler#installation)
+
+
+VPA has the following modes:
+
+* `Off` where right sizing recommendation will be generated but it won't change any existing pods
+* `Initial` when a new pod is created, it will be configured with the limits/requests that are considered appropriate
+* `Recreate` every time a new suitable sizing event happens (when better requests/limits are computed) pods may be evicted to apply the new configuration
+* `Auto` it is the same as `Recreate`, but in the future, it may support restart-free updates 
+
+By default, the VPA is configured to generate recommendations only. The following configuration enables it:
+```yaml
+verticalPodAutoscaler:
+  enabled: true
+```
+
+To see the recommendation you can run
+```bash
+~ $ kubectl get vpa 
+NAME            MODE   CPU   MEM   PROVIDED   AGE
+release-nginx   Auto                          24m
+
+
+~ $ kubectl describe vpa release-nginx
+```
+
+You can learn more about using Vertical Pod Autoscaler [the official
+documentation](https://github.com/kubernetes/autoscaler/tree/master/vertical-pod-autoscaler).
+
+back to [root README](/README.adoc#day-to-day-operations)

charts/k8s-service/templates/_capabilities_helpers.tpl

@@ -49,3 +49,12 @@
     {{- print "autoscaling/v2beta2" -}}
   {{- end -}}
 {{- end -}}
+
+{{/* Get VertialPodAutoscaler API Version */}}
+{{- define "gruntwork.verticalPodAutoscaler.apiVersion" -}}
+  {{- if and (.Capabilities.APIVersions.Has "autoscaling.k8s.io/v1") (semverCompare ">= 1.23-0" (include "gruntwork.kubeVersion" .)) -}}
+    {{- print "autoscaling.k8s.io/v1" -}}
+  {{- else -}}
+    {{- print "autoscaling.k8s.io/v1beta2" -}}
+  {{- end -}}
+{{- end -}}

charts/k8s-service/templates/verticalpodautoscaler.yaml

@@ -0,0 +1,40 @@
+{{- if .Values.verticalPodAutoscaler.enabled }}
+apiVersion: {{ include "gruntwork.verticalPodAutoscaler.apiVersion" . }}
+kind: VerticalPodAutoscaler
+metadata:
+  name: {{ include "k8s-service.fullname" . }}
+  namespace: {{ $.Release.Namespace }}
+spec:
+  targetRef:
+    apiVersion: apps/v1
+    kind: Deployment
+    name: {{ include "k8s-service.fullname" . }}
+  updatePolicy:
+    updateMode: {{ .Values.verticalPodAutoscaler.updateMode | quote }}
+    minReplicas: {{ .Values.verticalPodAutoscaler.minReplicas }}
+  resourcePolicy:
+    containerPolicies:
+    - containerName: {{ include "k8s-service.fullname" . }}
+      {{- if .Values.verticalPodAutoscaler.mainContainerResourcePolicy.minAllowed }}
+      minAllowed:
+      {{- toYaml .Values.verticalPodAutoscaler.mainContainerResourcePolicy.minAllowed | nindent 8 }}
+      {{- end }}
+
+      {{- if .Values.verticalPodAutoscaler.mainContainerResourcePolicy.maxAllowed }}
+      maxAllowed:
+      {{- toYaml .Values.verticalPodAutoscaler.mainContainerResourcePolicy.maxAllowed | nindent 8 }}
+      {{- end }}
+
+      {{- if .Values.verticalPodAutoscaler.mainContainerResourcePolicy.controlledResources }}
+      controlledResources: 
+      {{- toYaml .Values.verticalPodAutoscaler.mainContainerResourcePolicy.controlledResources | nindent 8 }}
+      {{- end }}
+
+      {{- if .Values.verticalPodAutoscaler.mainContainerResourcePolicy.controlledValues }}
+      controlledValues: {{ .Values.verticalPodAutoscaler.mainContainerResourcePolicy.controlledValues }}
+      {{- end }}
+
+    {{- if .Values.verticalPodAutoscaler.extraResourcePolicy }}
+    {{- toYaml .Values.verticalPodAutoscaler.extraResourcePolicy | nindent 4 }}
+    {{- end }}
+{{- end }}

charts/k8s-service/values.yaml

@@ -805,6 +805,159 @@ horizontalPodAutoscaler:
   minReplicas: 1
   maxReplicas: 10
 
+# verticalPodAutoscaler is a map that configures the Vertical Pod Autoscaler information for this pod
+# The expected keys of vpa are:
+#   - enabled                       (bool)   : Whether or not Vertical Pod Autoscaler should be created, if false the
+#                                              Vertical Pod Autoscaler will not be created
+#   - updateMode                    (string) : Controls when autoscaler applies changes to the pod resources. Possible values:
+#                                                * "Off" Changes are only recommended but the pods are not changed
+#                                                * "Initial" Changes are applied at pod creation only
+#                                                * "Recreate" Pods will be evicted to apply new scaling settings
+#                                                * "Auto" Currently it is equivalent with "Recreate" but in the future
+#                                                  it will update pods without eviction
+#   - minReplicas                   (int)    : Minimal number of replicas which need to be alive for Updater to attempt
+#                                              pod eviction (pending other checks like PDB).
+#   - evictionRequirements          (list)   : Requirements that need to be true for an eviction event to be attempted
+#   - mainContainerResourcePolicy   (object) : Configuration for the main container 
+#   - extraResourcePolicy           (list)   : Configuration injected. Can be used if other sideCarContainers are used
+#
+#
+# The expected attributes for objects of type evictionRequirement:
+#   - resource             (list)    :  A list of one or more resources that the condition applies to. These resources can be "cpu", "memory", etc.
+#   - changeRequirement    (string)  : The type of change required. Possible options:
+#                                      * "TargetHigherThanRequests" - The eviction is attempted when the new target is higher than the current requests,
+#                                        i.e., the Pod is scaled up.
+#                                      * "TargetLowerThanRequests" - The eviction is attempted if the new target is lower than the current requests,
+#                                        i.e., the Pod is scaled down.
+#
+# The expected attributes for `mainContainerResourcePolicy` object:
+#  - mode                    (string) : The mode of operation. Can be "Off", "Auto"
+#  - minAllowed              (object) : The minimum resources allowed for the container. 
+#  - maxAllowed              (object) : The maximum resources allowed for the container.
+#  - controlledResources     (list)   : The list of resources that the policy can control. Can include "cpu" and "memory".
+#  - controlledValues        (string) : The values that the policy can control. Can be "RequestsOnly" or "RequestsAndLimits".
+#
+#
+# The expected attributes for `extraResourcePolicy` list of objects:
+#  - containerName           (string) : The name of the container that the policy applies to.
+#  - mode                    (string) : The mode of operation. Can be "Off", "Auto"
+#  - minAllowed              (object) : The minimum resources allowed for the container.
+#  - maxAllowed              (object) : The maximum resources allowed for the container.
+#  - controlledResources     (list)   : The list of resources that the policy can control. Can include "cpu" and "memory".
+#  - controlledValues        (string) : The values that the policy can control. Can be "RequestsOnly" or "RequestsAndLimits".
+
+# The expected attributes for `minAllowed` and `maxAllowed` objects:
+#  - cpu (string): 
+#  - memory (string): 
+# 
+#
+# Example configuration that uses all the features
+#
+# ```yaml
+# verticalPodAutoscaler:
+#   enabled: true
+#   updateMode: "Auto"
+#   minPodsAvailable: 2
+#   evictionRequirements:
+#     # Allow scale up events if the CPU needs adjusting
+#     - resource: ["cpu"]  
+#       changeRequirement: TargetHigherThanRequests
+#
+#     # Allow scale up events if the Memory needs adjusting
+#     - resource: ["memory"]
+#       changeRequirement: TargetHigherThanRequests
+#
+#     # Allow scale down events if both the CPU and memory need adjusting
+#     - resource: ["cpu", "memory"]
+#       changeRequirement: TargetLowerThanRequests
+#   mainContainerResourcePolicy:
+#     # Enable scaling for the main container
+#     mode: "Auto"
+#     minAllowed:
+#       cpu: "0.2"
+#       memory: "200Mi"
+#     maxAllowed:
+#       cpu: "1"
+#       memory: "500Mi"
+#     # Scale both the CPU and Memory of the container
+#     controlledResources: ["cpu", "memory"]
+#     # Adjust both requests and limits
+#     controlledValues: "RequestsAndLimits"
+#   # Configuration for other containers (e.g. when sideCarContainers are used)
+#   extraResourcePolicy:
+#     # Disable the scaling for all other containers by default
+#     - containerName: "*"
+#       mode: "Off"
+#     # Assuming we have a sideCarContainers called "proxy"
+#     - containerName: "proxy"
+#       # Enable scaling actions
+#       mode: "Auto"
+#       minAllowed:
+#         cpu: "0.2"
+#       maxAllowed:
+#         cpu: "1"
+#       # Only scale the CPU
+#       controlledResources: ["cpu"]
+#       # Scale only the requests
+#       controlledValues: "RequestsOnly"
+# ```
+#
+#
+# This will result in the following configuration
+#
+# ```yaml
+# apiVersion: autoscaling.k8s.io/v1
+# kind: VerticalPodAutoscaler
+# metadata:
+#   name: XXXX
+# spec:
+#   resourcePolicy:
+#     containerPolicies:
+#     - containerName: XXXX
+#       controlledResources:
+#       - cpu
+#       - memory
+#       controlledValues: RequestsAndLimits
+#       maxAllowed:
+#         cpu: "1"
+#         memory: 500Mi
+#       minAllowed:
+#         cpu: "0.2"
+#         memory: 200Mi
+#     - containerName: ""
+#       mode: "Off"
+#     - containerName: proxy
+#       controlledResources:
+#       - cpu
+#       controlledValues: RequestsOnly
+#       maxAllowed:
+#         cpu: "1"
+#       minAllowed:
+#         cpu: "0.2"
+#       mode: Auto
+#   targetRef:
+#     apiVersion: apps/v1
+#     kind: Deployment
+#     name: m1-nginx
+#   updatePolicy:
+#     minReplicas: 2
+#     updateMode: Auto
+#```
+#
+# The default config will not create the Vertical Pod Autoscaler by setting enabled = false,
+# the default values are set so if enabled is true the verticalPodAutoscaler has valid values
+# and is running in recommendation only mode
+verticalPodAutoscaler:
+  enabled: false
+  updateMode: "Off"
+  minReplicas: 2
+  mainContainerResourcePolicy:
+    controlledResources: ["cpu", "memory"]
+    controlledValues: "RequestsAndLimits"
+  extraResourcePolicy:
+  - containerName: ""
+    mode: "Off"
+
 # customResources is a map that lets you define Kubernetes resources you want installed and configured as part of this chart.
 # The expected keys of customResources are:
 #   - enabled   (bool) : Whether or not the provided custom resource definitions should be created.