onepassword-operator/USAGEGUIDE.md

Usage Guide

Table of Contents

1. Configuration Options
2. Use Kubernetes Operator with Service Account
   • Create a Service Account
   • Create a Kubernetes secret
   • Deploy the Operator
3. Use Kubernetes Operator with Connect
   • Deploy with Helm
   • Deploy manually
4. Logging level
5. Usage examples
6. How 1Password Items Map to Kubernetes Secrets
7. Configuring Automatic Rolling Restarts of Deployments
8. Development

---

Configuration options

There are 2 ways 1Password Operator can talk to 1Password servers:

• 1Password Service Accounts
• 1Password Connect

---

Use Kubernetes Operator with Service Account

1. Create a service account

2. Create a Kubernetes secret for the Service Account

• Set OP_SERVICE_ACCOUNT_TOKEN environment variable to the service account token you created in the previous step. This token will be used by the operator to access 1Password items.
• Create Kubernetes secret:

kubectl create secret generic onepassword-service-account-token --from-literal=token="$OP_SERVICE_ACCOUNT_TOKEN"

3. Deploy the Operator

An sample Deployment yaml can be found at /config/manager/manager.yaml. To use Operator with Service Account, you need to set the OP_SERVICE_ACCOUNT_TOKEN environment variable in the /config/manager/manager.yaml. And remove OP_CONNECT_TOKEN and OP_CONNECT_HOST environment variables.

To further configure the 1Password Kubernetes Operator the following Environment variables can be set in the operator yaml:

• OP_SERVICE_ACCOUNT_TOKEN (required): Specifies Service Account token within Kubernetes to access the 1Password items.
• WATCH_NAMESPACE: (default: watch all namespaces): Comma separated list of what Namespaces to watch for changes.
• POLLING_INTERVAL (default: 600): The number of seconds the 1Password Kubernetes Operator will wait before checking for updates from 1Password.
• AUTO_RESTART (default: false): If set to true, the operator will restart any deployment using a secret from 1Password. This can be overwritten by namespace, deployment, or individual secret. More details on AUTO_RESTART can be found in the "Configuring Automatic Rolling Restarts of Deployments" section.

To deploy the operator, simply run the following command:

make deploy

Undeploy Operator

make undeploy

---

Use Kubernetes Operator with Connect

1. Deploy with Helm

2. Deploy manually

To further configure the 1Password Kubernetes Operator the following Environment variables can be set in the operator yaml:

• OP_CONNECT_HOST (required): Specifies the host name within Kubernetes in which to access the 1Password Connect.
• WATCH_NAMESPACE: (default: watch all namespaces): Comma separated list of what Namespaces to watch for changes.
• POLLING_INTERVAL (default: 600): The number of seconds the 1Password Kubernetes Operator will wait before checking for updates from 1Password Connect.
• MANAGE_CONNECT (default: false): If set to true, on deployment of the operator, a default configuration of the OnePassword Connect Service will be deployed to the current namespace.
• AUTO_RESTART (default: false): If set to true, the operator will restart any deployment using a secret from 1Password Connect. This can be overwritten by namespace, deployment, or individual secret. More details on AUTO_RESTART can be found in the "Configuring Automatic Rolling Restarts of Deployments" section.

---

Logging level

You can set the logging level by setting --zap-log-level as an arg on the containers to either debug, info or error. The default value is debug.

Example:

....
containers:
      - command:
        - /manager
        args:
        - --leader-elect
        - --zap-log-level=info
        image: 1password/onepassword-operator:latest
....

---

Usage examples

Find usage examples on 1Password developer documentation.

---

How 1Password Items Map to Kubernetes Secrets

The contents of the Kubernetes secret will be key-value pairs in which the keys are the fields of the 1Password item and the values are the corresponding values stored in 1Password. In case of fields that store files, the file's contents will be used as the value.

Within an item, if both a field storing a file and a field of another type have the same name, the file field will be ignored and the other field will take precedence.

Deleting the Deployment that you've created will automatically delete the created Kubernetes Secret only if the deployment is still annotated with operator.1password.io/item-path and operator.1password.io/item-name and no other deployment is using the secret.

If a 1Password Item that is linked to a Kubernetes Secret is updated within the POLLING_INTERVAL the associated Kubernetes Secret will be updated. However, if you do not want a specific secret to be updated you can add the tag operator.1password.io:ignore-secret to the item stored in 1Password. While this tag is in place, any updates made to an item will not trigger an update to the associated secret in Kubernetes.

If multiple 1Password vaults/items have the same title when using a title in the access path, the desired action will be performed on the oldest vault/item.

Titles and field names that include white space and other characters that are not a valid DNS subdomain name will create Kubernetes secrets that have titles and fields in the following format:

• Invalid characters before the first alphanumeric character and after the last alphanumeric character will be removed
• All whitespaces between words will be replaced by -
• All the letters will be lower-cased.

---

Configuring Automatic Rolling Restarts of Deployments

If a 1Password Item that is linked to a Kubernetes Secret is updated, any deployments configured to auto-restart AND are using that secret will be given a rolling restart the next time 1Password Connect is polled for updates.

There are many levels of granularity on which to configure auto restarts on deployments:

• Operator level
• Per-namespace
• Per-deployment

Operator Level: This method allows for managing auto restarts on all deployments within the namespaces watched by operator. Auto restarts can be enabled by setting the environment variable AUTO_RESTART to true. If the value is not set, the operator will default this value to false.

Per Namespace: This method allows for managing auto restarts on all deployments within a namespace. Auto restarts can by managed by setting the annotation operator.1password.io/auto-restart to either true or false on the desired namespace. An example of this is shown below:

# enabled auto restarts for all deployments within a namespace unless overwritten within a deployment
apiVersion: v1
kind: Namespace
metadata:
  name: "example-namespace"
  annotations:
    operator.1password.io/auto-restart: "true"

If the value is not set, the auto restart settings on the operator will be used. This value can be overwritten by deployment.

Per Deployment This method allows for managing auto restarts on a given deployment. Auto restarts can by managed by setting the annotation operator.1password.io/auto-restart to either true or false on the desired deployment. An example of this is shown below:

# enabled auto restarts for the deployment
apiVersion: v1
kind: Deployment
metadata:
  name: "example-deployment"
  annotations:
    operator.1password.io/auto-restart: "true"

If the value is not set, the auto restart settings on the namespace will be used.

Per OnePasswordItem Custom Resource This method allows for managing auto restarts on a given OnePasswordItem custom resource. Auto restarts can by managed by setting the annotation operator.1password.io/auto_restart to either true or false on the desired OnePasswordItem. An example of this is shown below:

# enabled auto restarts for the OnePasswordItem
apiVersion: onepassword.com/v1
kind: OnePasswordItem
metadata:
  name: example
  annotations:
    operator.1password.io/auto-restart: "true"

If the value is not set, the auto restart settings on the deployment will be used.

---

Development

How it works

This project aims to follow the Kubernetes Operator pattern

It uses Controllers which provides a reconcile function responsible for synchronizing resources until the desired state is reached on the cluster

Test It Out

1. Install the CRDs into the cluster:

make install

2. Run your controller (this will run in the foreground, so switch to a new terminal if you want to leave it running):

make run

NOTE: You can also run this in one step by running: make install run

Modifying the API definitions

If you are editing the API definitions, generate the manifests such as CRs or CRDs using:

make manifests

NOTE: Run make --help for more information on all potential make targets

More information can be found via the Kubebuilder Documentation