OpenSearch/x-pack/docs/en/security/get-started-security.asciidoc

377 lines
14 KiB
Plaintext

[role="xpack"]
[testenv="basic"]
[[security-getting-started]]
== Tutorial: Getting started with security
In this tutorial, you learn how to secure a cluster by configuring users and
roles in {es}, {kib}, {ls}, and {metricbeat}.
[float]
[[get-started-security-prerequisites]]
=== Before you begin
. Install and configure {es}, {kib}, {ls}, and {metricbeat} as described in
{stack-gs}/get-started-elastic-stack.html[Getting started with the {stack}].
+
--
IMPORTANT: To complete this tutorial, you must install the default {es} and
{kib} packages, which include role-based access control (RBAC) and native
authentication {security-features}. When you install these products, they apply
basic licenses with no expiration dates. All of the subsequent steps in this
tutorial assume that you are using a basic license. For more information, see
{subscriptions} and {kibana-ref}/managing-licenses.html[License management].
--
. Stop {ls}. The method for starting and stopping {ls} varies depending on whether
you are running it from the command line or running it as a service. For example,
if you are running {ls} from the command line, you can stop it by entering
`Ctrl-C`. See {logstash-ref}/shutdown.html[Shutting down {ls}].
. Stop {metricbeat}. For example, enter `Ctrl-C` on the command line where it is
running.
. Launch the {kib} web interface by pointing your browser to port 5601. For
example, http://127.0.0.1:5601[http://127.0.0.1:5601].
[role="xpack"]
[[get-started-enable-security]]
=== Enable {es} {security-features}
include::get-started-enable-security.asciidoc[]
. Enable single-node discovery in the `ES_PATH_CONF/elasticsearch.yml` file.
+
--
This tutorial involves a single node cluster, but if you had multiple
nodes, you would enable {es} {security-features} on every node in the cluster
and configure Transport Layer Security (TLS) for internode-communication, which
is beyond the scope of this tutorial. By enabling single-node discovery, we are
postponing the configuration of TLS. For example, add the following setting:
[source,yaml]
----
discovery.type: single-node
----
For more information, see
{ref}/bootstrap-checks.html#single-node-discovery[Single-node discovery].
--
When you enable {es} {security-features}, basic authentication is enabled by
default. To communicate with the cluster, you must specify a username and
password. Unless you <<anonymous-access,enable anonymous access>>, all requests
that don't include a user name and password are rejected.
[role="xpack"]
[[get-started-built-in-users]]
=== Create passwords for built-in users
include::get-started-builtin-users.asciidoc[]
You need these built-in users in subsequent steps, so choose passwords that you
can remember!
NOTE: This tutorial does not use the built-in `apm_system`, `logstash_system`,
`beats_system`, and `remote_monitoring_user` users, which are typically
associated with monitoring. For more information, see
{logstash-ref}/ls-security.html#ls-monitoring-user[Configuring credentials for {ls} monitoring]
and {metricbeat-ref}/monitoring.html[Monitoring {metricbeat}].
[role="xpack"]
[[get-started-kibana-user]]
=== Add the built-in user to {kib}
include::get-started-kibana-users.asciidoc[]
[role="xpack"]
[[get-started-authentication]]
=== Configure authentication
Now that you've set up the built-in users, you need to decide how you want to
manage all the other users.
The {stack} _authenticates_ users to ensure that they are valid. The
authentication process is handled by _realms_. You can use one or more built-in
realms, such as the native, file, LDAP, PKI, Active Directory, SAML, or Kerberos
realms. Alternatively, you can create your own custom realms. In this tutorial,
we'll use a native realm.
In general, you configure realms by adding `xpack.security.authc.realms`
settings in the `elasticsearch.yml` file. However, the native realm is available
by default when no other realms are configured. Therefore, you don't need to do
any extra configuration steps in this tutorial. You can jump straight to
creating users!
If you want to learn more about authentication and realms, see
<<setting-up-authentication>>.
[role="xpack"]
[[get-started-users]]
=== Create users
Let's create two users in the native realm.
. Log in to {kib} with the `elastic` built-in user.
. Go to the *Management / Security / Users* page:
+
--
[role="screenshot"]
image::security/images/management-builtin-users.jpg["User management screenshot in Kibana"]
In this example, you can see a list of built-in users.
--
. Click *Create new user*. For example, create a user for yourself:
+
--
[role="screenshot"]
image::security/images/create-user.jpg["Creating a user in Kibana"]
You'll notice that when you create a user, you can assign it a role. Don't
choose a role yet--we'll come back to that in subsequent steps.
--
. Click *Create new user* and create a `logstash_internal` user.
+
--
In {stack-gs}/get-started-elastic-stack.html[Getting started with the {stack}],
you configured {ls} to listen for {metricbeat}
input and to send the events to {es}. You therefore need to create a user
that {ls} can use to communicate with {es}. For example:
[role="screenshot"]
image::security/images/create-logstash-user.jpg["Creating a {ls} user in {kib}"]
--
[role="xpack"]
[[get-started-roles]]
=== Assign roles
By default, all users can change their own passwords, get information about
themselves, and run the `authenticate` API. If you want them to do more than
that, you need to give them one or more _roles_.
Each role defines a specific set of actions (such as read, create, or delete)
that can be performed on specific secured resources (such as data streams, indices, aliases,
documents, fields, or clusters). To help you get up and running, there are
built-in roles.
Go to the *Management / Security / Roles* page to see them:
[role="screenshot"]
image::security/images/management-roles.jpg["Role management screenshot in Kibana"]
Select a role to see more information about its privileges. For example, select
the `kibana_system` role to see its list of cluster and index privileges. To
learn more, see <<privileges-list-indices>>.
Let's assign the `kibana_admin` role to your user. Go back to the
*Management / Security / Users* page and select your user. Add the `kibana_admin`
role and save the change. For example:
[role="screenshot"]
image::security/images/assign-role.jpg["Assigning a role to a user in Kibana"]
This user now has administrative access to all features in {kib}.
For more information about granting access to Kibana see
{kibana-ref}/xpack-security-authorization.html[Kibana authorization].
If you completed all of the steps in
{stack-gs}/get-started-elastic-stack.html[Getting started with the {stack}], you should
have {metricbeat} data stored in {es}. Let's create two roles that grant
different levels of access to that data.
Go to the *Management / Security / Roles* page and click *Create role*.
Create a `metricbeat_reader` role that has `read` and `view_index_metadata`
privileges on the `metricbeat-*` indices:
[role="screenshot"]
image::security/images/create-reader-role.jpg["Creating a role in Kibana"]
Create a `metricbeat_writer` role that has `manage_index_templates` and `monitor`
cluster privileges, as well as `write`, `delete`, and `create_index` privileges
on the `metricbeat-*` indices:
[role="screenshot"]
image::security/images/create-writer-role.jpg["Creating another role in Kibana"]
Now go back to the *Management / Security / Users* page and assign these roles
to the appropriate users. Assign the `metricbeat_reader` role to your personal
user. Assign the `metricbeat_writer` role to the `logstash_internal` user.
The list of users should now contain all of the built-in users as well as the
two you created. It should also show the appropriate roles for your users:
[role="screenshot"]
image::security/images/management-users.jpg["User management screenshot in Kibana"]
If you want to learn more about authorization and roles, see <<authorization>>.
[role="xpack"]
[[get-started-logstash-user]]
=== Add user information in {ls}
In order for {ls} to send data successfully to {es}, you must configure its
authentication credentials in the {ls} configuration file.
. Configure {ls} to use the `logstash_internal` user and the password that you
created:
** If you don't mind having passwords visible in your configuration file, add
the following `user` and `password` settings in the `demo-metrics-pipeline.conf`
file in your {ls} directory:
+
--
[source,ruby]
----
...
output {
elasticsearch {
hosts => "localhost:9200"
manage_template => false
index => "%{[@metadata][beat]}-%{[@metadata][version]}-%{+YYYY.MM.dd}"
user => "logstash_internal" <1>
password => "your_password" <2>
}
}
----
<1> Specify the `logstash_internal` user that you created earlier in this tutorial.
<2> Specify the password that you chose for this user ID.
--
** If you prefer not to put your user ID and password in the configuration file,
store them in a keystore instead.
+
--
Run the following commands to create the {ls}
keystore and add the secure settings:
["source","sh",subs="attributes,callouts"]
----------------------------------------------------------------------
set +o history
export LOGSTASH_KEYSTORE_PASS=mypassword <1>
set -o history
./bin/logstash-keystore create
./bin/logstash-keystore add ES_USER
./bin/logstash-keystore add ES_PWD
----------------------------------------------------------------------
<1> You can optionally protect access to the {ls} keystore by storing a password
in an environment variable called `LOGSTASH_KEYSTORE_PASS`. For more information,
see {logstash-ref}/keystore.html#keystore-password[Keystore password].
When prompted, specify the `logstash_internal` user and its password for the
`ES_USER` and `ES_PWD` values.
NOTE: The {ls} keystore differs from the {kib} keystore. Whereas the {kib}
keystore enables you to store `kibana.yml` settings by name, the {ls} keystore
enables you to create arbitrary names that you can reference in the {ls}
configuration. To learn more, see
{logstash-ref}/keystore.html[Secrets keystore for secure settings].
You can now use these `ES_USER` and `ES_PWD` keys in your configuration
file. For example, add the `user` and `password` settings in the
`demo-metrics-pipeline.conf` file as follows:
[source,ruby]
----
...
output {
elasticsearch {
hosts => "localhost:9200"
manage_template => false
index => "%{[@metadata][beat]}-%{[@metadata][version]}-%{+YYYY.MM.dd}"
user => "${ES_USER}"
password => "${ES_PWD}"
}
}
----
--
. Start {ls} by using the appropriate method for your environment.
+
--
For example, to
run {ls} from a command line, go to the {ls} directory and enter the following
command:
["source","sh",subs="attributes,callouts"]
----------------------------------------------------------------------
./bin/logstash -f demo-metrics-pipeline.conf
----------------------------------------------------------------------
To start {ls} as a service, see
{logstash-ref}/running-logstash.html[Running {ls} as a service on Debian or RPM].
--
. If you were connecting directly from {metricbeat} to {es}, you would need
to configure authentication credentials for the {es} output in the {metricbeat}
configuration file. In
{stack-gs}/get-started-elastic-stack.html[Getting started with the {stack}],
however, you configured
{metricbeat} to send the data to {ls} for additional parsing, so no extra
settings are required in {metricbeat}. For more information, see
{metricbeat-ref}/securing-metricbeat.html[Securing {metricbeat}].
. Start {metricbeat} by using the appropriate method for your environment.
+
--
For example, on macOS, run the following command from the {metricbeat} directory:
["source","sh",subs="attributes,callouts"]
----------------------------------------------------------------------
./metricbeat -e
----------------------------------------------------------------------
For more methods, see {metricbeat-ref}/metricbeat-starting.html[Starting {metricbeat}].
--
Wait a few minutes for new data to be sent from {metricbeat} to {ls} and {es}.
[role="xpack"]
[[get-started-verify-users]]
=== View system metrics in {kib}
Log in to {kib} with the user ID that has `metricbeat_reader` and `kibana_user`
roles (for example, `jdoe`).
These roles enable the user to see the system metrics in {kib} (for example, on
the *Discover* page or in the
http://localhost:5601/app/kibana#/dashboard/Metricbeat-system-overview[{metricbeat} system overview dashboard]).
[float]
[[gs-security-nextsteps]]
=== What's next?
Congratulations! You've successfully set up authentication and authorization by
using the native realm. You learned how to create user IDs and roles that
prevent unauthorized access to the {stack}.
Later, when you're ready to increase the number of nodes in your cluster, you'll
want to encrypt communications across the {stack}. To learn how, read
<<encrypting-communications>>.
For more detailed information about securing the {stack}, see:
* {ref}/configuring-security.html[Configuring security in {es}]. Encrypt
inter-node communications, set passwords for the built-in users, and manage your
users and roles.
* {kibana-ref}/using-kibana-with-security.html[Configuring security in {kib}].
Set the authentication credentials in {kib} and encrypt communications between
the browser and the {kib} server.
* {logstash-ref}/ls-security.html[Configuring security in Logstash]. Set the
authentication credentials for Logstash and encrypt communications between
Logstash and {es}.
* Configuring security in the Beats. Configure authentication
credentials and encrypt connections to {es}. For example, see
{filebeat-ref}/securing-filebeat.html[Configure {filebeat} to use {security-features}].
* {hadoop-ref}/security.html[Configuring {es} for Apache Hadoop to use secured transport].