Using Kibana in a production environmentedit

How you deploy Kibana largely depends on your use case. If you are the only user, you can run Kibana on your local machine and configure it to point to whatever Elasticsearch instance you want to interact with. Conversely, if you have a large number of heavy Kibana users, you might need to load balance across multiple Kibana instances that are all connected to the same Elasticsearch instance.

While Kibana isn’t terribly resource intensive, we still recommend running Kibana separate from your Elasticsearch data or master nodes. To distribute Kibana traffic across the nodes in your Elasticsearch cluster, you can run Kibana and an Elasticsearch client node on the same machine. For more information, see Load balancing across multiple Elasticsearch nodes.

Using Elastic Stack security featuresedit

You can use Elastic Stack security features to control what Elasticsearch data users can access through Kibana.

When security features are enabled, Kibana users have to log in. They need to have the kibana_user role as well as access to the indices they will be working with in Kibana.

If a user loads a Kibana dashboard that accesses data in an index that they are not authorized to view, they get an error that indicates the index does not exist. The security features do not currently provide a way to control which users can load which dashboards.

For information about setting up Kibana users, see Configuring security in Kibana.

Require Content Security Policyedit

Kibana uses a Content Security Policy to help prevent the browser from allowing unsafe scripting, but older browsers will silently ignore this policy. If your organization does not need to support Internet Explorer 11 or much older versions of our other supported browsers, we recommend that you enable Kibana’s strict mode for content security policy, which will block access to Kibana for any browser that does not enforce even a rudimentary set of CSP protections.

To do this, set csp.strict to true in your kibana.yml:

csp.strict: true

Enabling SSLedit

See Encrypting communications.

Load Balancing Across Multiple Elasticsearch Nodesedit

If you have multiple nodes in your Elasticsearch cluster, the easiest way to distribute Kibana requests across the nodes is to run an Elasticsearch Coordinating only node on the same machine as Kibana. Elasticsearch Coordinating only nodes are essentially smart load balancers that are part of the cluster. They process incoming HTTP requests, redirect operations to the other nodes in the cluster as needed, and gather and return the results. For more information, see Node in the Elasticsearch reference.

To use a local client node to load balance Kibana requests:

  1. Install Elasticsearch on the same machine as Kibana.
  2. Configure the node as a Coordinating only node. In elasticsearch.yml, set node.data, node.master and node.ingest to false:

    # 3. You want this node to be neither master nor data node nor ingest node, but
    #    to act as a "search load balancer" (fetching data from nodes,
    #    aggregating results, etc.)
    #
    node.master: false
    node.data: false
    node.ingest: false
  3. Configure the client node to join your Elasticsearch cluster. In elasticsearch.yml, set the cluster.name to the name of your cluster.

    cluster.name: "my_cluster"
  4. Check your transport and HTTP host configs in elasticsearch.yml under network.host and transport.host. The transport.host needs to be on the network reachable to the cluster members, the network.host is the network for the HTTP connection for Kibana (localhost:9200 by default).

    network.host: localhost
    http.port: 9200
    
    # by default transport.host refers to network.host
    transport.host: <external ip>
    transport.tcp.port: 9300 - 9400
  5. Make sure Kibana is configured to point to your local client node. In kibana.yml, the elasticsearch.hosts setting should be set to ["localhost:9200"].

    # The Elasticsearch instance to use for all your queries.
    elasticsearch.hosts: ["http://localhost:9200"]

Load balancing across multiple Kibana instancesedit

To serve multiple Kibana installations behind a load balancer, you must change the configuration. See Configuring Kibana for details on each setting.

Settings unique across each Kibana instance:

server.uuid
server.name

Settings unique across each host (for example, running multiple installations on the same virtual machine):

logging.dest
path.data
pid.file
server.port

Settings that must be the same:

xpack.security.encryptionKey //decrypting session cookies
xpack.reporting.encryptionKey //decrypting reports stored in Elasticsearch

Separate configuration files can be used from the command line by using the -c flag:

bin/kibana -c config/instance1.yml
bin/kibana -c config/instance2.yml

High availability across multiple Elasticsearch nodesedit

Kibana can be configured to connect to multiple Elasticsearch nodes in the same cluster. In situations where a node becomes unavailable, Kibana will transparently connect to an available node and continue operating. Requests to available hosts will be routed in a round robin fashion.

Currently the Console application is limited to connecting to the first node listed.

In kibana.yml:

elasticsearch.hosts:
  - http://elasticsearch1:9200
  - http://elasticsearch2:9200

Related configurations include elasticsearch.sniffInterval, elasticsearch.sniffOnStart, and elasticsearch.sniffOnConnectionFault. These can be used to automatically update the list of hosts as a cluster is resized. Parameters can be found on the settings page.

Memoryedit

Kibana has a default maximum memory limit of 1.4 GB, and in most cases, we recommend leaving this unconfigured. In some scenarios, such as large reporting jobs, it may make sense to tweak limits to meet more specific requirements.

You can modify this limit by setting --max-old-space-size in the NODE_OPTIONS environment variable. For deb and rpm, packages this is passed in via /etc/default/kibana and can be appended to the bottom of the file.

The option accepts a limit in MB:

NODE_OPTIONS="--max-old-space-size=2048" bin/kibana