How to use Infisical to inject secrets into Kubernetes clusters.
The Infisical Secrets Operator is a Kubernetes controller that retrieves secrets from Infisical and stores them in a designated cluster.
It uses an InfisicalSecret resource to specify authentication and storage methods.
The operator continuously updates secrets and can also reload dependent deployments automatically.
If you are already using the External Secrets operator, you can view the integration documentation for it here.
# Example installing app version v0.2.0 and chart version 0.1.4
helm install --generate-name infisical-helm-charts/secrets-operator --version=0.1.4 --setcontrollerManager.manager.image.tag=v0.2.0
Once you have installed the operator to your cluster, you’ll need to create a InfisicalSecret custom resource definition (CRD).
example-infisical-secret-crd.yaml
apiVersion: secrets.infisical.com/v1alpha1
kind: InfisicalSecret
metadata:name: infisicalsecret-sample
labels:label-to-be-passed-to-managed-secret: sample-value
annotations:example.com/annotation-to-be-passed-to-managed-secret:"sample-value"spec:hostAPI: https://app.infisical.com/api
resyncInterval:10authentication:# Make sure to only have 1 authentication method defined, serviceToken/universalAuth.# If you have multiple authentication methods defined, it may cause issues.# (Deprecated) Service Token AuthserviceToken:serviceTokenSecretReference:secretName: service-token
secretNamespace: default
secretsScope:envSlug: <env-slug>secretsPath: <secrets-path>recursive:true# Universal AuthuniversalAuth:secretsScope:projectSlug: new-ob-em
envSlug: dev # "dev", "staging", "prod", etc..secretsPath:"/"# Root is "/"recursive:true# Wether or not to use recursive mode (Fetches all secrets in an environment from a given secret path, and all folders inside the path) / defaults to falsecredentialsRef:secretName: universal-auth-credentials
secretNamespace: default
# Native Kubernetes AuthkubernetesAuth:identityId: <machine-identity-id>serviceAccountRef:name: <service-account-name>namespace: <service-account-namespace># secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true# AWS IAM AuthawsIamAuth:identityId: <your-machine-identity-id># secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true# Azure AuthazureAuth:identityId: <your-machine-identity-id>resource: https://management.azure.com/&client_id=CLIENT_ID# (Optional) This is the Azure resource that you want to access. For example, "https://management.azure.com/". If no value is provided, it will default to "https://management.azure.com/"# secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true# GCP ID Token AuthgcpIdTokenAuth:identityId: <your-machine-identity-id># secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true# GCP IAM AuthgcpIamAuth:identityId: <your-machine-identity-id># secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:truemanagedSecretReference:secretName: managed-secret
secretNamespace: default
creationPolicy:"Orphan"## Owner | Orphan# secretType: kubernetes.io/dockerconfigjson
If you are fetching secrets from a self hosted instance of Infisical set the value of hostAPI to
https://your-self-hosted-instace.com/api
When hostAPI is not defined the operator fetches secrets from Infisical Cloud.
If you have installed your Infisical instance within the same cluster as the Infisical operator, you can optionally access the Infisical backend’s service directly without having to route through the public internet.
To achieve this, use the following address for the hostAPI field:
Make sure to replace <backend-svc-name> and <namespace> with the appropriate values for your backend service and namespace.
This property defines the time in seconds between each secret re-sync from
Infisical. Shorter time between re-syncs will require higher rate limits only
available on paid plans. Default re-sync interval is every 1 minute.
This block defines the method that will be used to authenticate with Infisical
so that secrets can be fetched
The universal machine identity authentication method is used to authenticate with Infisical. The client ID and client secret needs to be stored in a Kubernetes secret. This block defines the reference to the name and namespace of secret that stores these credentials.
Once you have created your machine identity and added it to your project(s), you will need to create a Kubernetes secret containing the identity credentials.
To quickly create a Kubernetes secret containing the identity credentials, you can run the command below.
Make sure you replace <your-identity-client-id> with the identity client ID and <your-identity-client-secret> with the identity client secret.
Add reference for the Kubernetes secret containing the identity credentials
Once the secret is created, add the secretName and secretNamespace of the secret that was just created under authentication.universalAuth.credentialsRef field in the InfisicalSecret resource.
Make sure to also populate the secretsScope field with the project slug
projectSlug, environment slug envSlug, and secrets path
secretsPath that you want to fetch secrets from. Please see the example
below.
Example
apiVersion: secrets.infisical.com/v1alpha1
kind: InfisicalSecret
metadata:name: infisicalsecret-sample-crd
spec:authentication:universalAuth:secretsScope:projectSlug: <project-slug># <-- project slugenvSlug: <env-slug># "dev", "staging", "prod", etc..secretsPath:"<secrets-path>"# Root is "/"credentialsRef:secretName: universal-auth-credentials # <-- name of the Kubernetes secret that stores our machine identity credentialssecretNamespace: default # <-- namespace of the Kubernetes secret that stores our machine identity credentials...
The Kubernetes machine identity authentication method is used to authenticate with Infisical. The identity ID is stored in a field in the InfisicalSecret resource. This authentication method can only be used within a Kubernetes environment.
1
Obtaining the token reviewer JWT for Infisical
1.1. Start by creating a service account in your Kubernetes cluster that will be used by Infisical to authenticate with the Kubernetes API Server.
1.2. Bind the service account to the system:auth-delegator cluster role. As described here, this role allows delegated authentication and authorization checks, specifically for Infisical to access the TokenReview API. You can apply the following configuration file:
1.3. Next, create a long-lived service account JWT token (i.e. the token reviewer JWT token) for the service account using this configuration file for a new Secret resource:
1.5. Finally, retrieve the token reviewer JWT token from the secret.
kubectl get secret infisical-auth-token -n default -o=jsonpath='{.data.token}'| base64 --decode
Keep this JWT token handy as you will need it for the Token Reviewer JWT field when configuring the Kubernetes Auth authentication method for the identity in step 2.
2
Creating an identity
To create an identity, head to your Organization Settings > Access Control > Machine Identities and press Create identity.
When creating an identity, you specify an organization level role for it to assume; you can configure roles in Organization Settings > Access Control > Organization Roles.
Now input a few details for your new identity. Here’s some guidance for each field:
Name (required): A friendly name for the identity.
Role (required): A role from the Organization Roles tab for the identity to assume. The organization role assigned will determine what organization level resources this identity can have access to.
Once you’ve created an identity, you’ll be prompted to configure the authentication method for it. Here, select Kubernetes Auth.
To learn more about each field of the Kubernetes native authentication method, see step 2 of guide.
3
Adding an identity to a project
To allow the operator to use the given identity to access secrets, you will need to add the identity to project(s) that you would like to grant it access to.
To do this, head over to the project you want to add the identity to and go to Project Settings > Access Control > Machine Identities and press Add identity.
Next, select the identity you want to add to the project and the project level role you want to allow it to assume. The project role assigned will determine what project level resources this identity can have access to.
4
Add your identity ID & service account to your InfisicalSecret resource
Once you have created your machine identity and added it to your project(s), you will need to add the identity ID to your InfisicalSecret resource.
In the authentication.kubernetesAuth.identityId field, add the identity ID of the machine identity you created.
See the example below for more details.
5
Add your Kubernetes service account token to the InfisicalSecret resource
Add the service account details from the previous steps under authentication.kubernetesAuth.serviceAccountRef.
Here you will need to enter the name and namespace of the service account.
The example below shows a complete InfisicalSecret resource with all required fields defined.
Make sure to also populate the secretsScope field with the project slug
projectSlug, environment slug envSlug, and secrets path
secretsPath that you want to fetch secrets from. Please see the example
below.
Example
example-kubernetes-auth.yaml
apiVersion: secrets.infisical.com/v1alpha1
kind: InfisicalSecret
metadata:name: infisicalsecret-sample-crd
spec:authentication:kubernetesAuth:identityId: <machine-identity-id>serviceAccountRef:name: <service-account-name>namespace: <service-account-namespace># secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true...
The AWS IAM machine identity authentication method is used to authenticate with Infisical. The identity ID is stored in a field in the InfisicalSecret resource. This authentication method can only be used within an AWS environment like an EC2 or a Lambda function.
Add your identity ID to your InfisicalSecret resource
Once you have created your machine identity and added it to your project(s), you will need to add the identity ID to your InfisicalSecret resource. In the authentication.awsIamAuth.identityId field, add the identity ID of the machine identity you created. See the example below for more details.
Make sure to also populate the secretsScope field with the project slug
projectSlug, environment slug envSlug, and secrets path
secretsPath that you want to fetch secrets from. Please see the example
below.
Example
example-aws-iam-auth.yaml
apiVersion: secrets.infisical.com/v1alpha1
kind: InfisicalSecret
metadata:name: infisicalsecret-sample-crd
spec:authentication:awsIamAuth:identityId: <your-machine-identity-id># secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true...
The Azure machine identity authentication method is used to authenticate with Infisical. The identity ID is stored in a field in the InfisicalSecret resource. This authentication method can only be used within an Azure environment.
Add your identity ID to your InfisicalSecret resource
Once you have created your machine identity and added it to your project(s), you will need to add the identity ID to your InfisicalSecret resource. In the authentication.azureAuth.identityId field, add the identity ID of the machine identity you created. See the example below for more details.
Make sure to also populate the secretsScope field with the project slug
projectSlug, environment slug envSlug, and secrets path
secretsPath that you want to fetch secrets from. Please see the example
below.
Example
example-azure-auth.yaml
apiVersion: secrets.infisical.com/v1alpha1
kind: InfisicalSecret
metadata:name: infisicalsecret-sample-crd
spec:authentication:azureAuth:identityId: <your-machine-identity-id># secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true...
The GCP ID Token machine identity authentication method is used to authenticate with Infisical. The identity ID is stored in a field in the InfisicalSecret resource. This authentication method can only be used within GCP environments.
Add your identity ID to your InfisicalSecret resource
Once you have created your machine identity and added it to your project(s), you will need to add the identity ID to your InfisicalSecret resource. In the authentication.gcpIdTokenAuth.identityId field, add the identity ID of the machine identity you created. See the example below for more details.
Make sure to also populate the secretsScope field with the project slug
projectSlug, environment slug envSlug, and secrets path
secretsPath that you want to fetch secrets from. Please see the example
below.
Example
example-gcp-id-token-auth.yaml
apiVersion: secrets.infisical.com/v1alpha1
kind: InfisicalSecret
metadata:name: infisicalsecret-sample-crd
spec:authentication:gcpIdTokenAuth:identityId: <your-machine-identity-id># secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true...
The GCP IAM machine identity authentication method is used to authenticate with Infisical. The identity ID is stored in a field in the InfisicalSecret resource. This authentication method can only be used both within and outside GCP environments.
Add your identity ID and service account token path to your InfisicalSecret resource
Once you have created your machine identity and added it to your project(s), you will need to add the identity ID to your InfisicalSecret resource. In the authentication.gcpIamAuth.identityId field, add the identity ID of the machine identity you created.
You’ll also need to add the service account key file path to your InfisicalSecret resource. In the authentication.gcpIamAuth.serviceAccountKeyFilePath field, add the path to your service account key file path. Please see the example below for more details.
Make sure to also populate the secretsScope field with the project slug
projectSlug, environment slug envSlug, and secrets path
secretsPath that you want to fetch secrets from. Please see the example
below.
Example
example-gcp-id-token-auth.yaml
apiVersion: secrets.infisical.com/v1alpha1
kind: InfisicalSecret
metadata:name: infisicalsecret-sample-crd
spec:authentication:gcpIamAuth:identityId: <your-machine-identity-id>serviceAccountKeyFilePath:"/path/to-service-account-key-file-path.json"# secretsScope is identical to the secrets scope in the universalAuth field in this sample.secretsScope:projectSlug: your-project-slug
envSlug: prod
secretsPath:"/path"recursive:true...
The service token required to authenticate with Infisical needs to be stored in a Kubernetes secret. This block defines the reference to the name and namespace of secret that stores this service token.
Follow the instructions below to create and store the service token in a Kubernetes secrets and reference it in your CRD.
1. Generate service token
You can generate a service token for an Infisical project by heading over to the Infisical dashboard then to Project Settings.
2. Create Kubernetes secret containing service token
Once you have generated the service token, you will need to create a Kubernetes secret containing the service token you generated.
To quickly create a Kubernetes secret containing the generated service token, you can run the command below. Make sure you replace <your-service-token-here> with your service token.
3. Add reference for the Kubernetes secret containing service token
Once the secret is created, add the name and namespace of the secret that was just created under authentication.serviceToken.serviceTokenSecretReference field in the InfisicalSecret resource.
Make sure to also populate the secretsScope field with the, environment slug
envSlug, and secrets path secretsPath that you want to fetch secrets
from. Please see the example below.
Example
apiVersion: secrets.infisical.com/v1alpha1
kind: InfisicalSecret
metadata:name: infisicalsecret-sample-crd
spec:authentication:serviceToken:serviceTokenSecretReference:secretName: service-token # <-- name of the Kubernetes secret that stores our service tokensecretNamespace: option # <-- namespace of the Kubernetes secret that stores our service tokensecretsScope:envSlug: <env-slug># "dev", "staging", "prod", etc..secretsPath: <secrets-path># Root is "/"...
The managedSecretReference field is used to define the target location for storing secrets retrieved from an Infisical project.
This field requires specifying both the name and namespace of the Kubernetes secret that will hold these secrets.
The Infisical operator will automatically create the Kubernetes secret with the specified name/namespace and keep it continuously updated.
Note: The managed secret be should be created in the same namespace as the deployment that will use it.
The name of the managed Kubernetes secret to be created
The namespace of the managed Kubernetes secret to be created.
Override the default Opaque type for managed secrets with this field. Useful for creating kubernetes.io/dockerconfigjson secrets.
Creation polices allow you to control whether or not owner references should be added to the managed Kubernetes secret that is generated by the Infisical operator.
This is useful for tools such as ArgoCD, where every resource requires an owner reference; otherwise, it will be pruned automatically.
Available options
Orphan (default)
Owner
When creation policy is set to Owner, the InfisicalSecret CRD must be in
the same namespace as where the managed kubernetes secret.
The operator will transfer all labels & annotations present on the InfisicalSecret CRD to the managed Kubernetes secret to be created.
Thus, if a specific label is required on the resulting secret, it can be applied as demonstrated in the following example:
Once you have configured the Infisical CRD with the required fields, you can apply it to your cluster.
After applying, you should notice that the managed secret has been created in the desired namespace your specified.
Incorporating the managed secret created by the operator into your deployment can be achieved through several methods.
Here, we will highlight three of the most common ways to utilize it. Learn more about Kubernetes secrets here
This will take all the secrets from your managed secret and expose them to your container
envFrom:-secretRef:name: managed-secret # managed secret name
```
Example usage in a deployment
```yaml
apiVersion: apps/v1
kind: Deployment
metadata:name: nginx-deployment
labels:app: nginx
spec:replicas:1selector:matchLabels:app: nginx
template:metadata:labels:app: nginx
spec:containers:-name: nginx
image: nginx:1.14.2
envFrom:-secretRef:name: managed-secret # <- name of managed secretports:-containerPort:80
This will allow you to select individual secrets by key name from your managed secret and expose them to your container
env:-name: SECRET_NAME # The environment variable's name which is made available in the containervalueFrom:secretKeyRef:name: managed-secret # managed secret namekey: SOME_SECRET_KEY # The name of the key which exists in the managed secret
Deployments using managed secrets don’t reload automatically on updates, so they may use outdated secrets unless manually redeployed.
To address this, we added functionality to automatically redeploy your deployment when its managed secret updates.
When a secret change occurs, the operator will check to see which deployments are using the operator-managed Kubernetes secret that received the update.
Then, for each deployment that has this annotation present, a rolling update will be triggered.
To configure global settings that will apply to all instances of InfisicalSecret, you can define these configurations in a Kubernetes ConfigMap.
For example, you can configure all InfisicalSecret instances to fetch secrets from a single backend API without specifying the hostAPI parameter for each instance.
All global configurations must reside in a Kubernetes ConfigMap named infisical-config in the namespace infisical-operator-system.
To apply global configuration to the operator, copy the following yaml into infisical-config.yaml file.
If the operator is unable to fetch secrets from the API, it will not affect the managed Kubernetes secret.
It will continue attempting to reconnect to the API indefinitely.
The InfisicalSecret resource uses the status.conditions field to report its current state and any errors encountered.
$ kubectl get infisicalSecrets
NAME AGE
infisicalsecret-sample 12s
$ kubectl describe infisicalSecret infisicalsecret-sample
...Spec:...Status:Conditions:Last Transition Time:2022-12-18T04:29:09ZMessage: Infisical controller has located the Infisical token in provided Kubernetes secret
Reason: OK
Status:TrueType: secrets.infisical.com/LoadedInfisicalToken
Last Transition Time:2022-12-18T04:29:10ZMessage:Failed to update secret because: 400 Bad Request
Reason: Error
Status:FalseType: secrets.infisical.com/ReadyToSyncSecrets
Events: <none>