Post

Run the cheapest Amazon EKS

Posted Updated
Preview Image
By Petr Ruzicka 20 min read

Sometimes it is necessary to save costs and run the Amazon EKS the “cheapest way”.

The following notes are about running Amazon EKS with lowest price.

Requirements:

Build Amazon EKS cluster

Requirements

If you would like to follow this documents and it’s task you will need to set up few environment variables like:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

# AWS Region
export AWS_DEFAULT_REGION="${AWS_DEFAULT_REGION:-us-east-1}"
# Hostname / FQDN definitions
export CLUSTER_FQDN="${CLUSTER_FQDN:-k01.k8s.mylabs.dev}"
# Base Domain: k8s.mylabs.dev
export BASE_DOMAIN="${CLUSTER_FQDN#*.}"
# Cluster Name: k01
export CLUSTER_NAME="${CLUSTER_FQDN%%.*}"
export MY_EMAIL="petr.ruzicka@gmail.com"
export TMP_DIR="${TMP_DIR:-${PWD}}"
export KUBECONFIG="${KUBECONFIG:-${TMP_DIR}/${CLUSTER_FQDN}/kubeconfig-${CLUSTER_NAME}.conf}"
# Tags used to tag the AWS resources
export TAGS="${TAGS:-Owner=${MY_EMAIL},Environment=dev,Cluster=${CLUSTER_FQDN}}"
AWS_ACCOUNT_ID=$(aws sts get-caller-identity --query "Account" --output text) && export AWS_ACCOUNT_ID
mkdir -pv "${TMP_DIR}/${CLUSTER_FQDN}"

You will need to configure AWS CLI and other secrets/variables.

1
2
3
4
5
6
7

# AWS Credentials
export AWS_ACCESS_KEY_ID="xxxxxxxxxxxxxxxxxx"
export AWS_SECRET_ACCESS_KEY="xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
export AWS_SESSION_TOKEN="xxxxxxxx"
export AWS_ROLE_TO_ASSUME="arn:aws:iam::7xxxxxxxxxx7:role/Gixxxxxxxxxxxxxxxxxxxxle"
export GOOGLE_CLIENT_ID="10xxxxxxxxxxxxxxxud.apps.googleusercontent.com"
export GOOGLE_CLIENT_SECRET="GOxxxxxxxxxxxxxxxtw"

Verify if all the necessary variables were set:

1
2
3
4
5
6
7
8

: "${AWS_ACCESS_KEY_ID?}"
: "${AWS_DEFAULT_REGION?}"
: "${AWS_SECRET_ACCESS_KEY?}"
: "${AWS_ROLE_TO_ASSUME?}"
: "${GOOGLE_CLIENT_ID?}"
: "${GOOGLE_CLIENT_SECRET?}"

echo -e "${MY_EMAIL} | ${CLUSTER_NAME} | ${BASE_DOMAIN} | ${CLUSTER_FQDN}\n${TAGS}"

Install necessary tools:

You can skip these steps if you have all the required software already installed.

Configure AWS Route 53 Domain delegation

DNS delegation steps should be done only once

Create DNS zone for EKS clusters:

1
2
3
4
5
6
7

export CLOUDFLARE_EMAIL="petr.ruzicka@gmail.com"
export CLOUDFLARE_API_KEY="1xxxxxxxxx0"

aws route53 create-hosted-zone --output json \
  --name "${BASE_DOMAIN}" \
  --caller-reference "$(date)" \
  --hosted-zone-config="{\"Comment\": \"Created by petr.ruzicka@gmail.com\", \"PrivateZone\": false}" | jq

Route53 k8s.mylabs.dev zone Route53 k8s.mylabs.dev zone

Use your domain registrar to change the nameservers for your zone (for example mylabs.dev) to use the Amazon Route 53 nameservers. Here is the way how you can find out the the Route 53 nameservers:

1
2
3
4

NEW_ZONE_ID=$(aws route53 list-hosted-zones --query "HostedZones[?Name==\`${BASE_DOMAIN}.\`].Id" --output text)
NEW_ZONE_NS=$(aws route53 get-hosted-zone --output json --id "${NEW_ZONE_ID}" --query "DelegationSet.NameServers")
NEW_ZONE_NS1=$(echo "${NEW_ZONE_NS}" | jq -r ".[0]")
NEW_ZONE_NS2=$(echo "${NEW_ZONE_NS}" | jq -r ".[1]")

Create the NS record in k8s.mylabs.dev (BASE_DOMAIN) for proper zone delegation. This step depends on your domain registrar - I’m using CloudFlare and using Ansible to automate it:

1
2

ansible -m cloudflare_dns -c local -i "localhost," localhost -a "zone=mylabs.dev record=${BASE_DOMAIN} type=NS value=${NEW_ZONE_NS1} solo=true proxied=no account_email=${CLOUDFLARE_EMAIL} account_api_token=${CLOUDFLARE_API_KEY}"
ansible -m cloudflare_dns -c local -i "localhost," localhost -a "zone=mylabs.dev record=${BASE_DOMAIN} type=NS value=${NEW_ZONE_NS2} solo=false proxied=no account_email=${CLOUDFLARE_EMAIL} account_api_token=${CLOUDFLARE_API_KEY}"

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56

localhost | CHANGED => {
    "ansible_facts": {
        "discovered_interpreter_python": "/usr/bin/python"
    },
    "changed": true,
    "result": {
        "record": {
            "content": "ns-885.awsdns-46.net",
            "created_on": "2020-11-13T06:25:32.18642Z",
            "id": "dxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxb",
            "locked": false,
            "meta": {
                "auto_added": false,
                "managed_by_apps": false,
                "managed_by_argo_tunnel": false,
                "source": "primary"
            },
            "modified_on": "2020-11-13T06:25:32.18642Z",
            "name": "k8s.mylabs.dev",
            "proxiable": false,
            "proxied": false,
            "ttl": 1,
            "type": "NS",
            "zone_id": "2xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxe",
            "zone_name": "mylabs.dev"
        }
    }
}
localhost | CHANGED => {
    "ansible_facts": {
        "discovered_interpreter_python": "/usr/bin/python"
    },
    "changed": true,
    "result": {
        "record": {
            "content": "ns-1692.awsdns-19.co.uk",
            "created_on": "2020-11-13T06:25:37.605605Z",
            "id": "9xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxb",
            "locked": false,
            "meta": {
                "auto_added": false,
                "managed_by_apps": false,
                "managed_by_argo_tunnel": false,
                "source": "primary"
            },
            "modified_on": "2020-11-13T06:25:37.605605Z",
            "name": "k8s.mylabs.dev",
            "proxiable": false,
            "proxied": false,
            "ttl": 1,
            "type": "NS",
            "zone_id": "2xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxe",
            "zone_name": "mylabs.dev"
        }
    }
}

CloudFlare mylabs.dev zone CloudFlare mylabs.dev zone

Create Route53

Create CloudFormation template containing Route53 zone.

Put new domain CLUSTER_FQDN to the Route 53 and configure the DNS delegation from the BASE_DOMAIN.

Create Route53 zone:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

tee "${TMP_DIR}/${CLUSTER_FQDN}/aws-cf-route53.yml" << \EOF
AWSTemplateFormatVersion: 2010-09-09
Description: Route53 entries

Parameters:
  BaseDomain:
    Description: "Base domain where cluster domains + their subdomains will live. Ex: k8s.mylabs.dev"
    Type: String
  ClusterFQDN:
    Description: "Cluster FQDN. (domain for all applications) Ex: k01.k8s.mylabs.dev"
    Type: String
Resources:
  HostedZone:
    Type: AWS::Route53::HostedZone
    Properties:
      Name: !Ref ClusterFQDN
  RecordSet:
    Type: AWS::Route53::RecordSet
    Properties:
      HostedZoneName: !Sub "${BaseDomain}."
      Name: !Ref ClusterFQDN
      Type: NS
      TTL: 60
      ResourceRecords: !GetAtt HostedZone.NameServers
EOF

if [[ $(aws cloudformation list-stacks --stack-status-filter CREATE_COMPLETE --query "StackSummaries[?starts_with(StackName, \`${CLUSTER_NAME}-route53\`) == \`true\`].StackName" --output text) == "" ]]; then
  # shellcheck disable=SC2001
  eval aws cloudformation create-stack \
    --parameters "ParameterKey=BaseDomain,ParameterValue=${BASE_DOMAIN} ParameterKey=ClusterFQDN,ParameterValue=${CLUSTER_FQDN}" \
    --stack-name "${CLUSTER_NAME}-route53" \
    --template-body "file://${TMP_DIR}/${CLUSTER_FQDN}/aws-cf-route53.yml" \
    --tags "$(echo "${TAGS}" | sed -e 's/\([^=]*\)=\([^,]*\),*/Key=\1,Value=\2 /g')" || true
fi

After running the CF stack you should see the following Route53 zones:

Route53 k01.k8s.mylabs.dev zone Route53 k01.k8s.mylabs.dev zone

Route53 k8s.mylabs.dev zone Route53 k8s.mylabs.dev zone

Create Amazon EKS

I’m going to use eksctl to create the Amazon EKS cluster.

eksctl

Create Amazon EKS using eksctl:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68

tee "${TMP_DIR}/${CLUSTER_FQDN}/eksctl-${CLUSTER_NAME}.yaml" << EOF
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
metadata:
  name: ${CLUSTER_NAME}
  region: ${AWS_DEFAULT_REGION}
  tags:
    karpenter.sh/discovery: ${CLUSTER_NAME}
    $(echo "${TAGS}" | sed "s/,/\\n    /g; s/=/: /g")
availabilityZones:
  - ${AWS_DEFAULT_REGION}a
  - ${AWS_DEFAULT_REGION}b
iam:
  withOIDC: true
  serviceAccounts:
    - metadata:
        name: cert-manager
        namespace: cert-manager
      wellKnownPolicies:
        certManager: true
      roleName: eksctl-${CLUSTER_NAME}-irsa-cert-manager
    - metadata:
        name: external-dns
        namespace: external-dns
      wellKnownPolicies:
        externalDNS: true
      roleName: eksctl-${CLUSTER_NAME}-irsa-external-dns
# Allow users which are consuming the AWS_ROLE_TO_ASSUME to access the EKS
iamIdentityMappings:
  - arn: arn:aws:iam::${AWS_ACCOUNT_ID}:role/admin
    groups:
      - system:masters
    username: admin
karpenter:
  # renovate: datasource=github-tags depName=aws/karpenter extractVersion=^(?<version>.*)$
  version: v0.31.4
  createServiceAccount: true
  withSpotInterruptionQueue: true
addons:
  - name: vpc-cni
    # min version 1.14.0
    version: latest
    configurationValues: |-
      enableNetworkPolicy: "true"
      env:
        ENABLE_PREFIX_DELEGATION: "true"
  - name: kube-proxy
  - name: coredns
  - name: aws-ebs-csi-driver
managedNodeGroups:
  - name: mng01-ng
    amiFamily: Bottlerocket
    # Minimal instance type for running add-ons + karpenter - ARM t4g.medium: 4.0 GiB, 2 vCPUs - 0.0336 hourly
    # Minimal instance type for running add-ons + karpenter - X86 t3a.medium: 4.0 GiB, 2 vCPUs - 0.0336 hourly
    instanceType: t4g.medium
    # Due to karpenter we need 2 instances
    desiredCapacity: 2
    availabilityZones:
      - ${AWS_DEFAULT_REGION}a
    minSize: 2
    maxSize: 5
    volumeSize: 20
    disablePodIMDS: true
    volumeEncrypted: true
    # For instances with less than 30 vCPUs the maximum number is 110 and for all other instances the maximum number is 250
    # https://docs.aws.amazon.com/eks/latest/userguide/cni-increase-ip-addresses.html
    maxPodsPerNode: 110
EOF

Get the kubeconfig to access the cluster:

1
2
3
4
5
6
7
8
9

if [[ ! -s "${KUBECONFIG}" ]]; then
  if ! eksctl get clusters --name="${CLUSTER_NAME}" &> /dev/null; then
    eksctl create cluster --config-file "${TMP_DIR}/${CLUSTER_FQDN}/eksctl-${CLUSTER_NAME}.yaml" --kubeconfig "${KUBECONFIG}"
  else
    eksctl utils write-kubeconfig --cluster="${CLUSTER_NAME}" --kubeconfig "${KUBECONFIG}"
  fi
fi

aws eks update-kubeconfig --name="${CLUSTER_NAME}"

Karpenter

Karpenter is a Kubernetes Node Autoscaler built for flexibility, performance, and simplicity.

Karpenter

Configure Karpenter:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

tee "${TMP_DIR}/${CLUSTER_FQDN}/k8s-karpenter-provisioner.yml" << EOF | kubectl apply -f -
apiVersion: karpenter.sh/v1alpha5
kind: Provisioner
metadata:
  name: default
spec:
  # Enables consolidation which attempts to reduce cluster cost by both removing
  # un-needed nodes and down-sizing those that can't be removed.
  # https://youtu.be/OB7IZolZk78?t=2629
  consolidation:
    enabled: true
  requirements:
    - key: karpenter.sh/capacity-type
      operator: In
      values: ["spot", "on-demand"]
    - key: kubernetes.io/arch
      operator: In
      values: ["amd64", "arm64"]
    - key: "topology.kubernetes.io/zone"
      operator: In
      values: ["${AWS_DEFAULT_REGION}a"]
    - key: karpenter.k8s.aws/instance-family
      operator: In
      values: ["t3a", "t4g"]
  kubeletConfiguration:
    maxPods: 110
  # Resource limits constrain the total size of the cluster.
  # Limits prevent Karpenter from creating new instances once the limit is exceeded.
  limits:
    resources:
      cpu: 8
      memory: 32Gi
  providerRef:
    name: default
  # Labels are arbitrary key-values that are applied to all nodes
  labels:
    managedBy: karpenter
    provisioner: default
---
apiVersion: karpenter.k8s.aws/v1alpha1
kind: AWSNodeTemplate
metadata:
  name: default
spec:
  amiFamily: Bottlerocket
  subnetSelector:
    karpenter.sh/discovery: ${CLUSTER_NAME}
  securityGroupSelector:
    karpenter.sh/discovery: ${CLUSTER_NAME}
  blockDeviceMappings:
    - deviceName: /dev/xvda
      ebs:
        volumeSize: 2Gi
        volumeType: gp3
        encrypted: true
    - deviceName: /dev/xvdb
      ebs:
        volumeSize: 20Gi
        volumeType: gp3
        encrypted: true
  tags:
    KarpenerProvisionerName: "default"
    Name: "${CLUSTER_NAME}-karpenter"
    $(echo "${TAGS}" | sed "s/,/\\n    /g; s/=/: /g")
EOF

aws-node-termination-handler

AWS Node Termination Handler gracefully handle EC2 instance shutdown within Kubernetes.

Install aws-node-termination-handler helm chart and modify the default values:

1
2
3
4
5
6
7
8

# renovate: datasource=helm depName=aws-node-termination-handler registryUrl=https://aws.github.io/eks-charts
AWS_NODE_TERMINATION_HANDLER_HELM_CHART_VERSION="0.21.0"

helm repo add eks https://aws.github.io/eks-charts/
tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-aws-node-termination-handler.yml" << EOF
awsRegion: ${AWS_DEFAULT_REGION}
EOF
helm upgrade --install --version "${AWS_NODE_TERMINATION_HANDLER_HELM_CHART_VERSION}" --namespace kube-system --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-aws-node-termination-handler.yml" aws-node-termination-handler eks/aws-node-termination-handler

mailhog

Mailhog will be used to receive email alerts form the Prometheus.

MailHog

Install mailhog helm chart and modify the default values.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

# renovate: datasource=helm depName=mailhog registryUrl=https://codecentric.github.io/helm-charts
MAILHOG_HELM_CHART_VERSION="5.2.3"

helm repo add codecentric https://codecentric.github.io/helm-charts
tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-mailhog.yml" << EOF
image:
  repository: docker.io/cd2team/mailhog
  tag: "1663459324"
ingress:
  enabled: true
  annotations:
    forecastle.stakater.com/expose: "true"
    forecastle.stakater.com/icon: https://raw.githubusercontent.com/sj26/mailcatcher/main/assets/images/logo_large.png
    forecastle.stakater.com/appName: Mailhog
    nginx.ingress.kubernetes.io/auth-url: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/auth
    nginx.ingress.kubernetes.io/auth-signin: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/start?rd=\$scheme://\$host\$request_uri
  ingressClassName: nginx
  hosts:
    - host: mailhog.${CLUSTER_FQDN}
      paths:
        - path: /
          pathType: ImplementationSpecific
  tls:
    - hosts:
        - mailhog.${CLUSTER_FQDN}
EOF
helm upgrade --install --version "${MAILHOG_HELM_CHART_VERSION}" --namespace mailhog --create-namespace --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-mailhog.yml" mailhog codecentric/mailhog

kube-prometheus-stack

kube-prometheus stack is a collection of Kubernetes manifests, Grafana dashboards, and Prometheus rules combined with documentation and scripts to provide easy to operate end-to-end Kubernetes cluster monitoring with Prometheus using the Prometheus Operator.

Prometheus

Install kube-prometheus-stack helm chart and modify the default values:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264

# renovate: datasource=helm depName=kube-prometheus-stack registryUrl=https://prometheus-community.github.io/helm-charts
KUBE_PROMETHEUS_STACK_HELM_CHART_VERSION="56.6.2"

helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-kube-prometheus-stack.yml" << EOF
defaultRules:
  rules:
    etcd: false
    kubernetesSystem: false
    kubeScheduler: false
alertmanager:
  config:
    global:
      smtp_smarthost: "mailhog.mailhog.svc.cluster.local:1025"
      smtp_from: "alertmanager@${CLUSTER_FQDN}"
    route:
      group_by: ["alertname", "job"]
      receiver: email-notifications
      routes:
        - receiver: email-notifications
          matchers: [ '{severity=~"warning|critical"}' ]
    receivers:
      - name: email-notifications
        email_configs:
          - to: "notification@${CLUSTER_FQDN}"
            require_tls: false
  ingress:
    enabled: true
    ingressClassName: nginx
    annotations:
      forecastle.stakater.com/expose: "true"
      forecastle.stakater.com/icon: https://raw.githubusercontent.com/stakater/ForecastleIcons/master/alert-manager.png
      forecastle.stakater.com/appName: Alert Manager
      nginx.ingress.kubernetes.io/auth-url: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/auth
      nginx.ingress.kubernetes.io/auth-signin: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/start?rd=\$scheme://\$host\$request_uri
    hosts:
      - alertmanager.${CLUSTER_FQDN}
    paths: ["/"]
    pathType: ImplementationSpecific
    tls:
      - hosts:
          - alertmanager.${CLUSTER_FQDN}
# https://github.com/grafana/helm-charts/blob/main/charts/grafana/values.yaml
grafana:
  defaultDashboardsEnabled: false
  serviceMonitor:
    enabled: true
  ingress:
    enabled: true
    ingressClassName: nginx
    annotations:
      forecastle.stakater.com/expose: "true"
      forecastle.stakater.com/icon: https://raw.githubusercontent.com/stakater/ForecastleIcons/master/grafana.png
      forecastle.stakater.com/appName: Grafana
      nginx.ingress.kubernetes.io/auth-url: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/auth
      nginx.ingress.kubernetes.io/auth-signin: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/start?rd=\$scheme://\$host\$request_uri
      nginx.ingress.kubernetes.io/configuration-snippet: |
        auth_request_set \$email \$upstream_http_x_auth_request_email;
        proxy_set_header X-Email \$email;
    hosts:
      - grafana.${CLUSTER_FQDN}
    paths: ["/"]
    pathType: ImplementationSpecific
    tls:
      - hosts:
          - grafana.${CLUSTER_FQDN}
  datasources:
    datasource.yaml:
      apiVersion: 1
      datasources:
        - name: Prometheus
          type: prometheus
          url: http://kube-prometheus-stack-prometheus.kube-prometheus-stack:9090/
          access: proxy
          isDefault: true
  dashboardProviders:
    dashboardproviders.yaml:
      apiVersion: 1
      providers:
        - name: "default"
          orgId: 1
          folder: ""
          type: file
          disableDeletion: false
          editable: true
          options:
            path: /var/lib/grafana/dashboards/default
  dashboards:
    default:
      1860-node-exporter-full:
        # renovate: depName="Node Exporter Full"
        gnetId: 1860
        revision: 33
        datasource: Prometheus
      3662-prometheus-2-0-overview:
        # renovate: depName="Prometheus 2.0 Overview"
        gnetId: 3662
        revision: 2
        datasource: Prometheus
      9852-stians-disk-graphs:
        # renovate: depName="node-exporter disk graphs"
        gnetId: 9852
        revision: 1
        datasource: Prometheus
      12006-kubernetes-apiserver:
        # renovate: depName="Kubernetes apiserver"
        gnetId: 12006
        revision: 1
        datasource: Prometheus
      9614-nginx-ingress-controller:
        # renovate: depName="NGINX Ingress controller"
        gnetId: 9614
        revision: 1
        datasource: Prometheus
      11875-kubernetes-ingress-nginx-eks:
        # renovate: depName="Kubernetes Ingress Nginx - EKS"
        gnetId: 11875
        revision: 1
        datasource: Prometheus
      15038-external-dns:
        # renovate: depName="External-dns"
        gnetId: 35038
        revision: 3
        datasource: Prometheus
      14314-kubernetes-nginx-ingress-controller-nextgen-devops-nirvana:
        # renovate: depName="Kubernetes Nginx Ingress Prometheus NextGen"
        gnetId: 14314
        revision: 2
        datasource: Prometheus
      13473-portefaix-kubernetes-cluster-overview:
        # renovate: depName="Portefaix / Kubernetes cluster Overview"
        gnetId: 13473
        revision: 2
        datasource: Prometheus
      # https://grafana.com/orgs/imrtfm/dashboards - https://github.com/dotdc/grafana-dashboards-kubernetes
      15760-kubernetes-views-pods:
        # renovate: depName="Kubernetes / Views / Pods"
        gnetId: 15760
        revision: 26
        datasource: Prometheus
      15757-kubernetes-views-global:
        # renovate: depName="Kubernetes / Views / Global"
        gnetId: 15757
        revision: 37
        datasource: Prometheus
      15758-kubernetes-views-namespaces:
        # renovate: depName="Kubernetes / Views / Namespaces"
        gnetId: 15758
        revision: 34
        datasource: Prometheus
      15759-kubernetes-views-nodes:
        # renovate: depName="Kubernetes / Views / Nodes"
        gnetId: 15759
        revision: 29
        datasource: Prometheus
      15761-kubernetes-system-api-server:
        # renovate: depName="Kubernetes / System / API Server"
        gnetId: 16761
        revision: 16
        datasource: Prometheus
      15762-kubernetes-system-coredns:
        # renovate: depName="Kubernetes / System / CoreDNS"
        gnetId: 15762
        revision: 17
        datasource: Prometheus
      19105-prometheus:
        # renovate: depName="Prometheus"
        gnetId: 29205
        revision: 2
        datasource: Prometheus
      16237-cluster-capacity:
        # renovate: depName="Cluster Capacity (Karpenter)"
        gnetId: 16237
        revision: 1
        datasource: Prometheus
      16236-pod-statistic:
        # renovate: depName="Pod Statistic (Karpenter)"
        gnetId: 16236
        revision: 1
        datasource: Prometheus
      19268-prometheus:
        # renovate: depName="Prometheus All Metrics"
        gnetId: 19268
        revision: 1
        datasource: Prometheus
  grafana.ini:
    analytics:
      check_for_updates: false
    server:
      root_url: https://grafana.${CLUSTER_FQDN}
    # Use oauth2-proxy instead of default Grafana Oauth
    auth.basic:
      enabled: false
    auth.proxy:
      enabled: true
      header_name: X-Email
      header_property: email
    users:
      auto_assign_org_role: Admin
  smtp:
    enabled: true
    host: "mailhog.mailhog.svc.cluster.local:1025"
    from_address: grafana@${CLUSTER_FQDN}
  networkPolicy:
    enabled: true
kubeControllerManager:
  enabled: false
kubeEtcd:
  enabled: false
kubeScheduler:
  enabled: false
kubeProxy:
  enabled: false
kube-state-metrics:
  networkPolicy:
    enabled: true
prometheus-node-exporter:
  networkPolicy:
    enabled: true
prometheusOperator:
  tls:
    enabled: false
  admissionWebhooks:
    enabled: false
  networkPolicy:
    enabled: true
prometheus:
  networkPolicy:
    enabled: false
  ingress:
    enabled: true
    ingressClassName: nginx
    annotations:
      forecastle.stakater.com/expose: "true"
      forecastle.stakater.com/icon: https://raw.githubusercontent.com/cncf/artwork/master/projects/prometheus/icon/color/prometheus-icon-color.svg
      forecastle.stakater.com/appName: Prometheus
      nginx.ingress.kubernetes.io/auth-url: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/auth
      nginx.ingress.kubernetes.io/auth-signin: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/start?rd=\$scheme://\$host\$request_uri
    paths: ["/"]
    pathType: ImplementationSpecific
    hosts:
      - prometheus.${CLUSTER_FQDN}
    tls:
      - hosts:
          - prometheus.${CLUSTER_FQDN}
  prometheusSpec:
    externalLabels:
      cluster: ${CLUSTER_FQDN}
    externalUrl: https://prometheus.${CLUSTER_FQDN}
    ruleSelectorNilUsesHelmValues: false
    serviceMonitorSelectorNilUsesHelmValues: false
    podMonitorSelectorNilUsesHelmValues: false
    retentionSize: 1GB
    walCompression: true
    storageSpec:
      volumeClaimTemplate:
        spec:
          storageClassName: gp2
          accessModes: ["ReadWriteOnce"]
          resources:
            requests:
              storage: 2Gi
EOF
helm upgrade --install --version "${KUBE_PROMETHEUS_STACK_HELM_CHART_VERSION}" --namespace kube-prometheus-stack --create-namespace --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-kube-prometheus-stack.yml" kube-prometheus-stack prometheus-community/kube-prometheus-stack

karpenter

Change karpenter default installation by upgrading: helm chart and modify the default values.

1
2
3
4
5
6
7
8
9
10
11
12
13

# renovate: datasource=github-tags depName=aws/karpenter extractVersion=^(?<version>.*)$
KARPENTER_HELM_CHART_VERSION="v0.31.4"

tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-karpenter.yml" << EOF
replicas: 1
serviceMonitor:
  enabled: true
settings:
  aws:
    enablePodENI: true
    reservedENIs: "1"
EOF
helm upgrade --install --version "${KARPENTER_HELM_CHART_VERSION}" --namespace karpenter --reuse-values --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-karpenter.yml" karpenter oci://public.ecr.aws/karpenter/karpenter

cert-manager

cert-manager adds certificates and certificate issuers as resource types in Kubernetes clusters, and simplifies the process of obtaining, renewing and using those certificates.

cert-manager

Install cert-manager helm chart and modify the default values. Service account cert-manager was created by eksctl.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

# renovate: datasource=helm depName=cert-manager registryUrl=https://charts.jetstack.io
CERT_MANAGER_HELM_CHART_VERSION="1.14.3"

helm repo add jetstack https://charts.jetstack.io
tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-cert-manager.yml" << EOF
installCRDs: true
serviceAccount:
  create: false
  name: cert-manager
extraArgs:
  - --cluster-resource-namespace=cert-manager
  - --enable-certificate-owner-ref=true
securityContext:
  fsGroup: 1001
prometheus:
  servicemonitor:
    enabled: true
webhook:
  networkPolicy:
    enabled: true
EOF
helm upgrade --install --version "${CERT_MANAGER_HELM_CHART_VERSION}" --namespace cert-manager --create-namespace --wait --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-cert-manager.yml" cert-manager jetstack/cert-manager

Add ClusterIssuers for Let’s Encrypt staging:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

tee "${TMP_DIR}/${CLUSTER_FQDN}/k8s-cert-manager-clusterissuer-staging.yml" << EOF | kubectl apply -f -
apiVersion: cert-manager.io/v1
kind: ClusterIssuer
metadata:
  name: letsencrypt-staging-dns
  namespace: cert-manager
  labels:
    letsencrypt: staging
spec:
  acme:
    server: https://acme-staging-v02.api.letsencrypt.org/directory
    email: ${MY_EMAIL}
    privateKeySecretRef:
      name: letsencrypt-staging-dns
    solvers:
      - selector:
          dnsZones:
            - ${CLUSTER_FQDN}
        dns01:
          route53:
            region: ${AWS_DEFAULT_REGION}
EOF

kubectl wait --namespace cert-manager --timeout=15m --for=condition=Ready clusterissuer --all

Create certificate:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

tee "${TMP_DIR}/${CLUSTER_FQDN}/k8s-cert-manager-certificate-staging.yml" << EOF | kubectl apply -f -
apiVersion: cert-manager.io/v1
kind: Certificate
metadata:
  name: ingress-cert-staging
  namespace: cert-manager
  labels:
    letsencrypt: staging
spec:
  secretName: ingress-cert-staging
  secretTemplate:
    labels:
      letsencrypt: staging
  issuerRef:
    name: letsencrypt-staging-dns
    kind: ClusterIssuer
  commonName: "*.${CLUSTER_FQDN}"
  dnsNames:
    - "*.${CLUSTER_FQDN}"
    - "${CLUSTER_FQDN}"
EOF

metrics-server

Metrics Server is a scalable, efficient source of container resource metrics for Kubernetes built-in autoscaling pipelines.

Install metrics-server helm chart and modify the default values:

1
2
3
4
5
6
7
8
9
10
11

# renovate: datasource=helm depName=metrics-server registryUrl=https://kubernetes-sigs.github.io/metrics-server/
METRICS_SERVER_HELM_CHART_VERSION="3.12.0"

helm repo add metrics-server https://kubernetes-sigs.github.io/metrics-server/
tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-metrics-server.yml" << EOF
metrics:
  enabled: true
serviceMonitor:
  enabled: true
EOF
helm upgrade --install --version "${METRICS_SERVER_HELM_CHART_VERSION}" --namespace kube-system --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-metrics-server.yml" metrics-server metrics-server/metrics-server

external-dns

ExternalDNS synchronizes exposed Kubernetes Services and Ingresses with DNS providers.

ExternalDNS

Install external-dns helm chart and modify the default values. external-dns will take care about DNS records. Service account external-dns was created by eksctl.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

# renovate: datasource=helm depName=external-dns registryUrl=https://kubernetes-sigs.github.io/external-dns/
EXTERNAL_DNS_HELM_CHART_VERSION="1.14.3"

helm repo add external-dns https://kubernetes-sigs.github.io/external-dns/
tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-external-dns.yml" << EOF
domainFilters:
  - ${CLUSTER_FQDN}
interval: 20s
policy: sync
serviceAccount:
  create: false
  name: external-dns
serviceMonitor:
  enabled: true
EOF
helm upgrade --install --version "${EXTERNAL_DNS_HELM_CHART_VERSION}" --namespace external-dns --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-external-dns.yml" external-dns external-dns/external-dns

ingress-nginx

ingress-nginx is an Ingress controller for Kubernetes using NGINX as a reverse proxy and load balancer.

Install ingress-nginx helm chart and modify the default values.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60

# renovate: datasource=helm depName=ingress-nginx registryUrl=https://kubernetes.github.io/ingress-nginx
INGRESS_NGINX_HELM_CHART_VERSION="4.9.1"

kubectl wait --namespace cert-manager --for=condition=Ready --timeout=10m certificate ingress-cert-staging

helm repo add ingress-nginx https://kubernetes.github.io/ingress-nginx
tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-ingress-nginx.yml" << EOF
controller:
  allowSnippetAnnotations: true
  ingressClassResource:
    default: true
  admissionWebhooks:
    networkPolicyEnabled: true
  extraArgs:
    default-ssl-certificate: "cert-manager/ingress-cert-staging"
  service:
    annotations:
      service.beta.kubernetes.io/aws-load-balancer-type: nlb
      service.beta.kubernetes.io/aws-load-balancer-additional-resource-tags: ${TAGS//\'/}
  metrics:
    enabled: true
    serviceMonitor:
      enabled: true
    prometheusRule:
      enabled: true
      rules:
        - alert: NGINXConfigFailed
          expr: count(nginx_ingress_controller_config_last_reload_successful == 0) > 0
          for: 1s
          labels:
            severity: critical
          annotations:
            description: bad ingress config - nginx config test failed
            summary: uninstall the latest ingress changes to allow config reloads to resume
        - alert: NGINXCertificateExpiry
          expr: (avg(nginx_ingress_controller_ssl_expire_time_seconds) by (host) - time()) < 604800
          for: 1s
          labels:
            severity: critical
          annotations:
            description: ssl certificate(s) will expire in less then a week
            summary: renew expiring certificates to avoid downtime
        - alert: NGINXTooMany500s
          expr: 100 * ( sum( nginx_ingress_controller_requests{status=~"5.+"} ) / sum(nginx_ingress_controller_requests) ) > 5
          for: 1m
          labels:
            severity: warning
          annotations:
            description: Too many 5XXs
            summary: More than 5% of all requests returned 5XX, this requires your attention
        - alert: NGINXTooMany400s
          expr: 100 * ( sum( nginx_ingress_controller_requests{status=~"4.+"} ) / sum(nginx_ingress_controller_requests) ) > 5
          for: 1m
          labels:
            severity: warning
          annotations:
            description: Too many 4XXs
            summary: More than 5% of all requests returned 4XX, this requires your attention
EOF
helm upgrade --install --version "${INGRESS_NGINX_HELM_CHART_VERSION}" --namespace ingress-nginx --create-namespace --wait --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-ingress-nginx.yml" ingress-nginx ingress-nginx/ingress-nginx

forecastle

Forecastle is a control panel which dynamically discovers and provides a launchpad to access applications deployed on Kubernetes.

Forecastle

Install forecastle helm chart and modify the default values.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

# renovate: datasource=helm depName=forecastle registryUrl=https://stakater.github.io/stakater-charts
FORECASTLE_HELM_CHART_VERSION="1.0.136"

helm repo add stakater https://stakater.github.io/stakater-charts
tee "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-forecastle.yml" << EOF
forecastle:
  config:
    namespaceSelector:
      any: true
    title: Launch Pad
  networkPolicy:
    enabled: true
  ingress:
    enabled: true
    annotations:
      nginx.ingress.kubernetes.io/auth-signin: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/start?rd=\$scheme://\$host\$request_uri
      nginx.ingress.kubernetes.io/auth-url: https://oauth2-proxy.${CLUSTER_FQDN}/oauth2/auth
    className: nginx
    hosts:
      - host: ${CLUSTER_FQDN}
        paths:
          - path: /
            pathType: Prefix
    tls:
      - hosts:
          - ${CLUSTER_FQDN}
EOF
helm upgrade --install --version "${FORECASTLE_HELM_CHART_VERSION}" --namespace forecastle --create-namespace --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-forecastle.yml" forecastle stakater/forecastle

oauth2-proxy

Use oauth2-proxy to protect the endpoints by Google Authentication.

OAuth2 Proxy

Install oauth2-proxy helm chart and modify the default values.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

# renovate: datasource=helm depName=oauth2-proxy registryUrl=https://oauth2-proxy.github.io/manifests
OAUTH2_PROXY_HELM_CHART_VERSION="6.24.1"

helm repo add oauth2-proxy https://oauth2-proxy.github.io/manifests
cat > "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-oauth2-proxy.yml" << EOF
config:
  clientID: ${GOOGLE_CLIENT_ID}
  clientSecret: ${GOOGLE_CLIENT_SECRET}
  cookieSecret: "$(openssl rand -base64 32 | head -c 32 | base64)"
  configFile: |-
    cookie_domains = ".${CLUSTER_FQDN}"
    set_authorization_header = "true"
    set_xauthrequest = "true"
    upstreams = [ "file:///dev/null" ]
    whitelist_domains = ".${CLUSTER_FQDN}"
authenticatedEmailsFile:
  enabled: true
  restricted_access: |-
    ${MY_EMAIL}
ingress:
  enabled: true
  className: nginx
  hosts:
    - oauth2-proxy.${CLUSTER_FQDN}
  tls:
    - hosts:
        - oauth2-proxy.${CLUSTER_FQDN}
metrics:
  servicemonitor:
    enabled: true
EOF
helm upgrade --install --version "${OAUTH2_PROXY_HELM_CHART_VERSION}" --namespace oauth2-proxy --create-namespace --values "${TMP_DIR}/${CLUSTER_FQDN}/helm_values-oauth2-proxy.yml" oauth2-proxy oauth2-proxy/oauth2-proxy

Clean-up

Clean-up

Remove EKS cluster and created components:

1
2
3

if eksctl get cluster --name="${CLUSTER_NAME}"; then
  eksctl delete cluster --name="${CLUSTER_NAME}" --force
fi

Remove Route 53 DNS records from DNS Zone:

1
2
3
4
5
6
7
8
9
10

CLUSTER_FQDN_ZONE_ID=$(aws route53 list-hosted-zones --query "HostedZones[?Name==\`${CLUSTER_FQDN}.\`].Id" --output text)
if [[ -n "${CLUSTER_FQDN_ZONE_ID}" ]]; then
  aws route53 list-resource-record-sets --hosted-zone-id "${CLUSTER_FQDN_ZONE_ID}" | jq -c '.ResourceRecordSets[] | select (.Type != "SOA" and .Type != "NS")' |
    while read -r RESOURCERECORDSET; do
      aws route53 change-resource-record-sets \
        --hosted-zone-id "${CLUSTER_FQDN_ZONE_ID}" \
        --change-batch '{"Changes":[{"Action":"DELETE","ResourceRecordSet": '"${RESOURCERECORDSET}"' }]}' \
        --output text --query 'ChangeInfo.Id'
    done
fi

Remove orphan EC2s created by Karpenter:

1
2
3
4

for EC2 in $(aws ec2 describe-instances --filters "Name=tag:kubernetes.io/cluster/${CLUSTER_NAME},Values=owned" Name=instance-state-name,Values=running --query "Reservations[].Instances[].InstanceId" --output text) ; do
  echo "Removing EC2: ${EC2}"
  aws ec2 terminate-instances --instance-ids "${EC2}"
done

Remove CloudWatch log group:

1

aws logs delete-log-group --log-group-name "/aws/eks/${CLUSTER_NAME}/cluster"

Remove CloudFormation stack:

1

aws cloudformation delete-stack --stack-name "${CLUSTER_NAME}-route53"

Wait for all CloudFormation stacks to be deleted:

1
2

aws cloudformation wait stack-delete-complete --stack-name "${CLUSTER_NAME}-route53"
aws cloudformation wait stack-delete-complete --stack-name "eksctl-${CLUSTER_NAME}-cluster"

Remove Volumes and Snapshots related to the cluster (just in case):

1
2
3
4

for VOLUME in $(aws ec2 describe-volumes --filter "Name=tag:KubernetesCluster,Values=${CLUSTER_NAME}" "Name=tag:kubernetes.io/cluster/${CLUSTER_NAME},Values=owned" --query 'Volumes[].VolumeId' --output text) ; do
  echo "*** Removing Volume: ${VOLUME}"
  aws ec2 delete-volume --volume-id "${VOLUME}"
done

Remove ${TMP_DIR}/${CLUSTER_FQDN} directory:

1

[[ -d "${TMP_DIR}/${CLUSTER_FQDN}" ]] && rm -rf "${TMP_DIR}/${CLUSTER_FQDN}" && [[ -d "${TMP_DIR}" ]] && rmdir "${TMP_DIR}" || true

Enjoy … 😉

Kubernetes, Amazon EKS
Amazon EKS k8s kubernetes karpenter eksctl cert-manager external-dns podinfo
This post is licensed under CC BY 4.0 by the author.