IMPORTANT: No additional bug fixes or documentation updates will be released for this version. For the latest information, see the current release documentation.
« Operator privileges Operator-only functionality »
Elastic Docs ›Elasticsearch Guide [8.15] ›Secure the Elastic Stack ›Operator privileges

Configure operator privileges

edit

Configure operator privileges

edit

This feature is designed for indirect use by Elasticsearch Service, Elastic Cloud Enterprise, and Elastic Cloud on Kubernetes. Direct use is not supported.

Before you can use operator privileges, you must enable the feature on all nodes in the cluster and designate operator users.

Enable operator privileges

edit

In order to use the operator privileges feature, it must be enabled explicitly on each node in the cluster. Add the following setting in each elasticsearch.yml file:

xpack.security.operator_privileges.enabled: true

If the node is already running before you make this change, you must restart the node for the feature to take effect.

The feature needs to be either enabled or disabled consistently across all nodes in a cluster. Otherwise, you can get inconsistent behaviour depending on which node first receives a request and which node executes it.

When operator privileges are enabled on a cluster, specific functionalities are restricted and can be executed only by users who have been explicitly designated as operator users. If a regular user attempts to execute these functionalities (even if they have the superuser role), a security exception occurs.

Designate operator users

edit

Operator users are just normal Elasticsearch users with special rights to perform operator-only functionalities. They are specified in an operator_users.yml file, which is located in the config directory (as defined by the ES_PATH_CONF environment variable). Similar to other security config files, the operator_users.yml file is local to a node and does not apply globally to the cluster. This means, in most cases, the same file should be distributed or copied to all nodes in a cluster.

The operator_users.yml file defines a set of criteria that an authenticating user must match to be considered as an operator user. The following snippet shows an example of such a file:

operator: 
  - usernames: ["system_agent_1","system_agent_2"] 
    realm_type: "file" 
    auth_type: "realm"

A fixed value of operator signals the beginning of the definition.

A list of user names allowed for operator users. This field is mandatory.

The type of the authenticating realm allowed for operator users. The default and only acceptable value is file.

The authentication type allowed for operator users. The default and only acceptable value is realm.

You must specify at least the usernames field. If no other fields are specified, their default values are used. All fields must be matched for a user to be qualified as an operator user. You can also specify multiple groups of criteria. This is currently not very useful since this feature does not yet support other realms or authentication types.

There are also two implicit rules that affect which users are operator users:

  1. If the authenticating user runs as another user, neither of them are considered to be operator users.
  2. All Internal users are implicitly operator users.

After a user is designated as an operator user, they are still subject to regular RBAC user authorization checks. That is to say, in addition to specifying that a user is an operator user, you must also grant them the necessary Elasticsearch roles to perform their tasks. Consequently, it is entirely possible that an operator user can encounter an "access denied" error and fail to perform certain actions due to RBAC check failures. In short, an operator user is not automatically a superuser.

« Operator privileges Operator-only functionality »