Parameter Store

Parameter Store

A ParameterStore points to AWS SSM Parameter Store in a certain account within a defined region. You should define Roles that define fine-grained access to individual secrets and pass them to ESO using spec.provider.aws.role. This way users of the SecretStore can only access the secrets necessary.

apiVersion: external-secrets.io/v1alpha1
kind: SecretStore
metadata:
  name: secretstore-sample
spec:
  provider:
    aws:
      service: ParameterStore
      # define a specific role to limit access
      # to certain secrets
      role: iam-role
      region: eu-central-1
      auth:
        secretRef:
          accessKeyIDSecretRef:
            name: awssm-secret
            key: access-key
          secretAccessKeySecretRef:
            name: awssm-secret
            key: secret-access-key

NOTE: In case of a ClusterSecretStore, Be sure to provide namespace in accessKeyIDSecretRef and secretAccessKeySecretRef with the namespaces where the secrets reside.

API Pricing & Throttling

The SSM Parameter Store API is charged by throughput and is available in different tiers, see pricing. Please estimate your costs before using ESO. Cost depends on the RefreshInterval of your ExternalSecrets.

IAM Policy

Create a IAM Policy to pin down access to secrets matching dev-*, for further information see AWS Documentation:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "ssm:GetParameter*"
      ],
      "Resource": "arn:aws:ssm:us-east-2:123456789012:parameter/dev-*"
    }
  ]
}

JSON Secret Values

You can store JSON objects in a parameter. You can access nested values or arrays using gjson syntax:

Consider the following JSON object that is stored in the Parameter Store key my-json-secret:

{
  "name": {"first": "Tom", "last": "Anderson"},
  "friends": [
    {"first": "Dale", "last": "Murphy"},
    {"first": "Roger", "last": "Craig"},
    {"first": "Jane", "last": "Murphy"}
  ]
}

This is an example on how you would look up nested keys in the above json object:

apiVersion: external-secrets.io/v1alpha1
kind: ExternalSecret
metadata:
  name: example
spec:
  # [omitted for brevity]
  data:
  - secretKey: firstname
    remoteRef:
      key: my-json-secret
      property: name.first # Tom
  - secretKey: first_friend
    remoteRef:
      key: my-json-secret
      property: friends.1.first # Roger

AWS Authentication

Controller's Pod Identity

This is basicially a zero-configuration authentication method that inherits the credentials from the runtime environment using the aws sdk default credential chain.

You can attach a role to the pod using IRSA, kiam or kube2iam. When no other authentication method is configured in the Kind=Secretstore this role is used to make all API calls against AWS Secrets Manager or SSM Parameter Store.

Based on the Pod's identity you can do a sts:assumeRole before fetching the secrets to limit access to certain keys in your provider. This is optional.

apiVersion: external-secrets.io/v1alpha1
kind: SecretStore
metadata:
  name: team-b-store
spec:
  provider:
    aws:
      service: SecretsManager
      region: eu-central-1
      # optional: do a sts:assumeRole before fetching secrets
      role: team-b

Access Key ID & Secret Access Key

You can store Access Key ID & Secret Access Key in a Kind=Secret and reference it from a SecretStore.

apiVersion: external-secrets.io/v1alpha1
kind: SecretStore
metadata:
  name: team-b-store
spec:
  provider:
    aws:
      service: SecretsManager
      region: eu-central-1
      # optional: assume role before fetching secrets
      role: team-b
      auth:
        secretRef:
          accessKeyIDSecretRef:
            name: awssm-secret
            key: access-key
          secretAccessKeySecretRef:
            name: awssm-secret
            key: secret-access-key

NOTE: In case of a ClusterSecretStore, Be sure to provide namespace in accessKeyIDSecretRef, secretAccessKeySecretRef with the namespaces where the secrets reside.

EKS Service Account credentials

This feature lets you use short-lived service account tokens to authenticate with AWS. You must have Service Account Volume Projection enabled - it is by default on EKS. See EKS guide on how to set up IAM roles for service accounts.

The big advantage of this approach is that ESO runs without any credentials.

apiVersion: v1
kind: ServiceAccount
metadata:
  annotations:
    eks.amazonaws.com/role-arn: arn:aws:iam::123456789012:role/team-a
  name: my-serviceaccount
  namespace: default

Reference the service account from above in the Secret Store:

apiVersion: external-secrets.io/v1alpha1
kind: SecretStore
metadata:
  name: secretstore-sample
spec:
  provider:
    aws:
      service: SecretsManager
      region: eu-central-1
      auth:
        jwt:
          serviceAccountRef:
            name: my-serviceaccount

NOTE: In case of a ClusterSecretStore, Be sure to provide namespace for serviceAccountRef with the namespace where the service account resides.

Custom Endpoints

You can define custom AWS endpoints if you want to use regional, vpc or custom endpoints. See List of endpoints for Secrets Manager, Secure Systems Manager and Security Token Service.

Use the following environment variables to point the controller to your custom endpoints. Note: All resources managed by this controller are affected.

ENV VAR	DESCRIPTION
AWS_SECRETSMANAGER_ENDPOINT	Endpoint for the Secrets Manager Service. The controller uses this endpoint to fetch secrets from AWS Secrets Manager.
AWS_SSM_ENDPOINT	Endpoint for the AWS Secure Systems Manager. The controller uses this endpoint to fetch secrets from SSM Parameter Store.
AWS_STS_ENDPOINT	Endpoint for the Security Token Service. The controller uses this endpoint when creating a session and when doing assumeRole or assumeRoleWithWebIdentity calls.