README.md

Deploy helm chart

Introduction

This chart bootstraps a deployment on a Kubernetes cluster using Helm package manager.

Prerequisites

The code was tested on Kubernetes cluster (v1.21.11) with Istio (1.12.6)

Istio is installed with Istio Ingress Gateway

• Kubernetes cluster version can be checked with the command:

  kubectl version --short | grep Server

  The output will be similar to the following:

  Server Version: v1.21.11-gke.1100

• Istio version can be checked in different ways, it is out of scope for this README. You can find more information here.

  The following command shows how to check version if Anthos Service Mesh is used:

  kubectl -n istio-system get pods -lapp=istiod -o=jsonpath='{.items[0].metadata.labels.istio\.io/rev}'

  The output will be similar to the following:

  asm-1132-5

It is possible to use other versions, but it hasn't been tested

This example describes installation in Development mode:

• In this mode helm chart is installed to the namespace not labeled with Istio.

  More information about labeling can be found here (Istio) or here (Anthos Service Mesh)

  You can find all labels for your namespace with the command:

  kubectl get namespace <namespace> -o jsonpath={.metadata.labels}

  The output shows that there are no any labels related to Istio:

  {"kubernetes.io/metadata.name":"default"}

  When the namespace is labeled with Istio, the output could be:

  {"istio-injection":"enabled","kubernetes.io/metadata.name":"default"}

Operation system

The code works in Debian-based Linux (Debian 10 and Ubuntu 20.04) and Windows WSL 2. Also, it works but is not guaranteed in Google Cloud Shell. All other operating systems, including macOS, are not verified and supported.

Packages

Packages are only needed for installation from a local computer.

• HELM (version: v3.7.1 or higher) helm

  Helm version can be checked with the command:

  helm version --short

  The output will be similar to the following:

  v3.7.1+gd141386

• Kubectl (version: v1.21.0 or higher) kubectl

  Kubectl version can be checked with the command:

  kubectl version --short | grep Client

  The output will be similar to the following:

  Client Version: v1.21.0

Installation

First you need to set variables in values.yaml file using any code editor. Some of the values are prefilled, but you need to specify some values as well. You can find more information about them below.

Global variables

Name	Description	Type	Default	Required
global.domain	your domain	string	-	yes
global.limitsEnabled	whether CPU and memory limits are enabled	boolean	true	yes
global.dataPartitionId	data partition id	string	-	yes
global.logLevel	severity of logging level	string	ERROR	yes
global.tier	Only PROD must be used to enable autoscaling	string	""	no
global.autoscaling	enables horizontal pod autoscaling, when tier=PROD	boolean	true	yes

Configmap variables

Name	Description	Type	Default	Required
data.logLevel	logging severity level for this service only	string	-	yes, only if differs from the global.logLevel
data.projectId	your Google Cloud project id	string	-	yes
data.dataProjectId	in case of multiproject cloud installation (services and data stored in different project) the name of data project	string	-	only in case of multiproject installation
data.partitionHost	partition host	string	partition	yes
data.partitionNamespace	datastore namespace where partition will store the data	string	partition	yes
data.datafierSa	datafier service account	string	datafier	yes
data.indexerAugmenterEnabled	enable indexer Augmenter	string	false	no
data.elasticHttps	Configuration of the communication protocol for search and indexer services	boolean	-	yes

Deployment variables

Name	Description	Type	Default	Required
data.requestsCpu	amount of requests CPU	string	10m	yes
data.requestsMemory	amount of requests memory	string	400Mi	yes
data.limitsCpu	CPU limit	string	500m	only if global.limitsEnabled is true
data.limitsMemory	memory limit	string	1G	only if global.limitsEnabled is true
data.serviceAccountName	name of your service account	string	partition	yes
data.image	path to the image in a registry	string	-	yes
data.imagePullPolicy	when to pull the image	string	IfNotPresent	yes
data.bootstrapImage	name of the bootstrap image	string	-	yes

Configuration variables

Name	Description	Type	Default	Required
conf.appName	name of the app	string	partition	yes
conf.configmap	configmap to be used	string	partition-config	yes

ISTIO variables

Name	Description	Type	Default	Required
istio.proxyCPU	CPU request for Envoy sidecars	string	10m	yes
istio.proxyCPULimit	CPU limit for Envoy sidecars	string	500m	yes
istio.proxyMemory	memory request for Envoy sidecars	string	100Mi	yes
istio.proxyMemoryLimit	memory limit for Envoy sidecars	string	512Mi	yes
istio.bootstrapProxyCPU	CPU request for Envoy sidecars	string	10m	yes
istio.bootstrapProxyCPULimit	CPU limit for Envoy sidecars	string	100m	yes
istio.sidecarInject	whether Istio sidecar will be injected. Setting to false reduces security, because disables authorization policy.	boolean	true	yes
istio.partitionEditors	list of users who can edit partitions. If it is empty, the service allows all external GET requests and POST/PUT/PATCH for bootstrap k8s service account. It accepts gc accounts only	list	-	no

Horizontal Pod Autoscaling (HPA) variables (works only if tier=PROD and autoscaling=true)

Name	Description	Type	Default	Required
hpa.minReplicas	minimum number of replicas	integer	3	only if global.autoscaling is true and global.tier is PROD
hpa.maxReplicas	maximum number of replicas	integer	15	only if global.autoscaling is true and global.tier is PROD
hpa.targetType	type of measurements: AverageValue or Value	string	"AverageValue"	only if global.autoscaling is true and global.tier is PROD
hpa.targetValue	threshold value to trigger the scaling up	integer	400	only if global.autoscaling is true and global.tier is PROD
hpa.behaviorScaleUpStabilizationWindowSeconds	time to start implementing the scale up when it is triggered	integer	10	only if global.autoscaling is true and global.tier is PROD
hpa.behaviorScaleUpPoliciesValue	the maximum number of new replicas to create (in percents from current state)	integer	50	only if global.autoscaling is true and global.tier is PROD
hpa.behaviorScaleUpPoliciesPeriodSeconds	pause for every new scale up decision	integer	15	only if global.autoscaling is true and global.tier is PROD
hpa.behaviorScaleDownStabilizationWindowSeconds	time to start implementing the scale down when it is triggered	integer	60	only if global.autoscaling is true and global.tier is PROD
hpa.behaviorScaleDownPoliciesValue	the maximum number of replicas to destroy (in percents from current state)	integer	25	only if global.autoscaling is true and global.tier is PROD
hpa.behaviorScaleDownPoliciesPeriodSeconds	pause for every new scale down decision	integer	60	only if global.autoscaling is true and global.tier is PROD

Limits variables

Name	Description	Type	Default	Required
limits.maxTokens	maximum number of requests per fillInterval	integer	300	only if global.autoscaling is true and global.tier is PROD
limits.tokensPerFill	number of new tokens allowed every fillInterval	integer	300	only if global.autoscaling is true and global.tier is PROD
limits.fillInterval	time interval	string	"1s"	only if global.autoscaling is true and global.tier is PROD

Methodology for Parameter Calculation variables: hpa.targetValue, limits.maxTokens and limits.tokensPerFill

The parameters hpa.targetValue, limits.maxTokens and limits.tokensPerFill were determined through empirical testing during load testing. These tests were conducted using the N2D machine series, which can run on either AMD EPYC Milan or AMD EPYC Rome processors. The values were fine-tuned to ensure optimal performance under typical workloads.

Recommendations for New Instance Types

When changing the instance type to a newer generation, such as the C3D series, it is essential to conduct new load testing. This ensures the parameters are recalibrated to match the performance characteristics of the new processor architecture, optimizing resource utilization and maintaining application stability.

Install the helm chart

Run this command from within this directory:

helm install gc-partition-deploy .

Uninstalling the Chart

To uninstall the helm deployment:

helm uninstall gc-partition-deploy

To delete secrets and PVCs:

kubectl delete secret --all; kubectl delete pvc --all

Move-to-Top