Kubernetes => Route53 Mapping Service

This is a Kubernetes service that polls services and ingresses (in all namespaces) that are configured with the label dns=route53 and adds the appropriate alias to the domain specified by the annotation domainName=sub.mydomain.io. Multiple domains and top level domains are also supported: domainName=.mydomain.io,sub1.mydomain.io,sub2.mydomain.io

Usage

route53-kubernetes ReplicationController

The following is an example ReplicationController definition for route53-kubernetes:

Create the ReplicationController via kubectl create -f <name_of_route53-kubernetes-rc.yaml>

Note: We don't currently sign our docker images. So, please use our images at your own risk.

apiVersion: v1
kind: ReplicationController
metadata:
  name: route53-kubernetes
  namespace: kube-system
  labels:
    app: route53-kubernetes
spec:
  replicas: 1
  selector:
    app: route53-kubernetes
  template:
    metadata:
      labels:
        app: route53-kubernetes
    spec:
      containers:
        - image: quay.io/molecule/route53-kubernetes:v1.3.0
          name: route53-kubernetes
          env:
            - name: INGRESS_SERVICE_SELECTOR
              value: ingress=endpoint
            - name: DNS_RECORD_TYPE
              value: CNAME
            - name: DNS_RECORD_TTL
              value: 300

This service expects that it's running on a Kubernetes node on AWS and that the IAM profile for that node is set up to allow the following, along with the default permissions needed by Kubernetes:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": "route53:ListHostedZonesByName",
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": "elasticloadbalancing:DescribeLoadBalancers",
            "Resource": "*"
        },
        {
            "Effect": "Allow",
            "Action": "route53:ChangeResourceRecordSets",
            "Resource": "*"
        }
    ]
}

Ingress support

Service support ingress k8s resources. Because nginx ingress controller create on service for all ingress resources we use selector (configurable with INGRESS_SERVICE_SELECTOR environment variable) to find valid k8s service.

Default selector: ingress=endpoint

DNS resource type

Service support "A" and "CNAME" record types. By default service create "A" record for each domain record. To specify default record type use DNS_RECORD_TYPE environment variable. Each k8s service \ ingress can override dns record type by annotation.

DNS resource TTL

Service support TTL (in seconds) configuration for dns records. Default value is 300 seconds. Default value can be overridden with DNS_RECORD_TTL environment variable. You can use annotation dnsRecordType to provide service \ ingress specific ttl value.

Service Configuration

Given the following Kubernetes service definition:

apiVersion: v1
kind: Service
metadata:
  name: my-app
  labels:
    app: my-app
    role: web
    dns: route53
  annotations:
    domainName: "test.mydomain.com"
    dnsRecordType: "CNAME"
    dnsRecordTTL: 600
spec:
  selector:
    app: my-app
    role: web
  ports:
  - name: web
    port: 80
    protocol: TCP
    targetPort: web
  - name: web-ssl
    port: 443
    protocol: TCP
    targetPort: web-ssl
  type: LoadBalancer

A "CNAME" record for test.mydomain.com will be created which points to the ELB that is configured by kubernetes. This assumes that a hosted zone exists in Route53 for mydomain.com. Any record that previously existed for that dns record will be updated.

dnsRecordType and dnsRecordTTL annotations are optional.

Ingress Configuration

Given the following Kubernetes ingress definition:

apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: test-ingress
  labels:
    dns: route53
  annotations:
    domainName: "test.mydomain.com"
    dnsRecordType: "A"
    dnsRecordTTL: 300
spec:
  backend:
    serviceName: testsvc
    servicePort: 80

An "A" record for test.mydomain.com will be created as an alias to the ELB that is used by ingress controller service by kubernetes. This assumes that a hosted zone exists in Route53 for mydomain.com. Any record that previously existed for that dns record will be updated.

dnsRecordType and dnsRecordTTL annotations are optional.

Alternative setup

This setup shows some alternative ways to configure route53-kubernetes. First, you can specify kubernetes certs manually if you do not have service accounts enabled. Second, access to AWS can be configured through a Shared Credentials File.

apiVersion: v1
kind: ReplicationController
metadata:
  name: route53-kubernetes
  namespace: kube-system
  labels:
    app: route53-kubernetes
spec:
  replicas: 1
  selector:
    app: route53-kubernetes
  template:
    metadata:
      labels:
        app: route53-kubernetes
    spec:
      volumes:
        - name: ssl-cert
          secret:
            secretName: kube-ssl
        - name: aws-creds
          secret:
            secretName: aws-creds
      containers:
        - image: quay.io/molecule/route53-kubernetes:v1.3.0
          name: route53-kubernetes
          volumeMounts:
            - name: ssl-cert
              mountPath: /opt/certs
              readOnly: true
            - name: aws-creds
              mountPath: /opt/creds
              readOnly: true
          env:
            - name: "CA_FILE_PATH"
              value: "/opt/certs/ca.pem"
            - name: "CERT_FILE_PATH"
              value: "/opt/certs/cert.pem"
            - name: "KEY_FILE_PATH"
              value: "/opt/certs/key.pem"
            - name: "AWS_SHARED_CREDENTIALS_FILE"
              value: "/opt/creds/credentials"

Building locally

Install dependencies

We use glide to manage dependencies. To fetch the dependencies to your local vendor/ folder please run:

glide install -v

Build the Image

You may choose to use Docker images for route53-kubernetes on our Quay namespace or to build the binary, docker image, and push the docker image from scratch. See the Makefile for more information on doing this process manually.