Alerting and action settings in Kibana

A string of 32 or more characters used to encrypt sensitive properties on alerting rules and actions before they’re stored in Elasticsearch. Third party credentials — such as the username and password used to connect to an SMTP service — are an example of encrypted properties.

Kibana offers a CLI tool to help generate this encryption key.

If not set, Kibana will generate a random key on startup, but all alerting and action functions will be blocked. Generated keys are not allowed for alerting and actions because when a new key is generated on restart, existing encrypted data becomes inaccessible. For the same reason, alerting and actions in high-availability deployments of Kibana will behave unexpectedly if the key isn’t the same on all instances of Kibana.

Although the key can be specified in clear text in kibana.yml, it’s recommended to store this key securely in the Kibana Keystore. Be sure to back up the encryption key value somewhere safe, as your alerting rules and actions will cease to function due to decryption failures should you lose it. If you want to rotate the encryption key, be sure to follow the instructions on encryption key rotation.

A list of hostnames that Kibana is allowed to connect to when built-in actions are triggered. It defaults to ["*"], allowing any host, but keep in mind the potential for SSRF attacks when hosts are not explicitly added to the allowed hosts. An empty list [] can be used to block built-in actions from making any external connections.

Note that hosts associated with built-in actions, such as Slack and PagerDuty, are not automatically added to allowed hosts. If you are not using the default ["*"] setting, you must ensure that the corresponding endpoints are added to the allowed hosts as well.

A list of custom host settings to override existing global settings. Default: an empty list.

Each entry in the list must have a url property, to associate a connection type (mail or https), hostname and port with the remaining options in the entry.

In the following example, two custom host settings are defined. The first provides a custom host setting for mail server mail.example.com using port 465 that supplies server certificate authentication data from both a file and inline, and requires TLS for the connection. The second provides a custom host setting for https server webhook.example.com which turns off server certificate authentication, that will allow Kibana to connect to the server if it’s using a self-signed certificate. The individual properties that can be used in the settings are documented below.

xpack.actions.customHostSettings:
  - url: smtp://mail.example.com:465
    ssl:
      verificationMode: 'full'
      certificateAuthoritiesFiles: [ 'one.crt' ]
      certificateAuthoritiesData: |
          -----BEGIN CERTIFICATE-----
          MIIDTD...
          CwUAMD...
          ... multiple lines of certificate data ...
          -----END CERTIFICATE-----
    smtp:
      requireTLS: true
  - url: https://webhook.example.com
    ssl:
      verificationMode: 'none'

The settings in xpack.actions.customHostSettings can be used to override the global option xpack.actions.ssl.verificationMode and provide customized TLS settings on a per-server basis. Set xpack.actions.ssl.verificationMode to the value to be used by default for all servers, then add an entry in xpack.actions.customHostSettings for every server that requires customized settings.

A URL associated with this custom host setting. Should be in the form of protocol://hostname:port, where protocol is https or smtp. If the port is not provided, 443 is used for https and 25 is used for smtp. The smtp URLs are used for the Email actions that use this server, and the https URLs are used for actions which use https to connect to services.

Entries with https URLs can use the ssl options, and entries with smtp URLs can use both the ssl and smtp options.

No other URL values should be part of this URL, including paths, query strings, and authentication information. When an http or smtp request is made as part of running an action, only the protocol, hostname, and port of the URL for that request are used to look up these configuration values.

A list of allowed email domains which can be used with the email connector. When this setting is not used, all email domains are allowed. When this setting is used, if any email is attempted to be sent that (a) includes an addressee with an email domain that is not in the allowlist, or (b) includes a from address domain that is not in the allowlist, it will fail with a message indicating the email is not allowed.

A list of action types that are enabled. It defaults to ["*"], enabling all types. The names for built-in Kibana action types are prefixed with a . and include: .email, .index, .jira, .opsgenie, .pagerduty, .resilient, .server-log, .servicenow, .servicenow-itom, .servicenow-sir, .slack, .swimlane, .teams, .thehive, .tines, .torq, .xmatters, .gen-ai, .bedrock, .gemini, .d3security, and .webhook. An empty list [] will disable all action types.

Disabled action types will not appear as an option when creating new connectors, but existing connectors and actions of that type will remain in Kibana and will not function.

Specifies the proxy URL to use, if using a proxy for actions. By default, no proxy is used.

Proxies may be used to proxy http or https requests through a proxy using the http or https protocol. Kibana only uses proxies in "CONNECT" mode (sometimes referred to as "tunneling" TCP mode, compared to HTTP mode). That is, Kibana will always make requests through a proxy using the HTTP CONNECT method.

If your proxy is using the https protocol (vs the http protocol), the setting xpack.actions.ssl.proxyVerificationMode: none will likely be needed, unless your proxy’s certificates are signed using a publicly available certificate authority.

There is currently no support for using basic authentication with a proxy (authentication for the proxy itself, not the URL being requested through the proxy).

To help diagnose problems using a proxy, you can use the curl command with options to use your proxy, and log debug information, with the following command, replacing the proxy and target URLs as appropriate. This will force the request to be made to the proxy in tunneling mode, and display some of the interaction between the client and the proxy.

curl --verbose --proxytunnel --proxy http://localhost:8080 http://example.com

Specifies hostnames which should not use the proxy, if using a proxy for actions. The value is an array of hostnames as strings. Example:

# If applicable, include the subdomain in the hostname
xpack.actions.proxyBypassHosts: [ "events.pagerduty.com" ]

By default, all hosts will use the proxy, but if an action’s hostname is in this list, the proxy will not be used. The settings xpack.actions.proxyBypassHosts and xpack.actions.proxyOnlyHosts cannot be used at the same time.

Specifies hostnames which should only use the proxy, if using a proxy for actions. The value is an array of hostnames as strings. Example:

# If applicable, include the subdomain in the hostname
xpack.actions.proxyOnlyHosts: [ "events.pagerduty.com" ]

By default, no hosts will use the proxy, but if an action’s hostname is in this list, the proxy will be used. The settings xpack.actions.proxyBypassHosts and xpack.actions.proxyOnlyHosts cannot be used at the same time.

Controls the verification for the server certificate that Elastic Maps Server receives when making an outbound SSL/TLS connection for actions. Valid values are full, certificate, and none. Use full to perform hostname verification, certificate to skip hostname verification, and none to skip verification. Default: full. Equivalent Kibana setting.

This setting can be overridden for specific URLs by using the setting xpack.actions.customHostSettings[n].ssl.verificationMode (described above) to a different value.

Overrides the configs under xpack.actions.run for the connector type with the given ID. List the connector type identifier and its settings in an array of objects. For example:

xpack.actions.run:
    maxAttempts: 1
    connectorTypeOverrides:
        - id: '.server-log'
          maxAttempts: 5

A configuration URL that varies by connector:

An application ID that varies by connector:

A client identifier that varies by connector:

For a Webhook - Case Management connector, specifies a stringified JSON payload with Mustache variables that is sent to the create comment URL to create a case comment. The required variable is case.description.

For a Webhook - Case Management connector, specifies a REST API URL string to create a case comment by ID in the third-party system.

For a Webhook - Case Management connector, specifies a stringified JSON payload with Mustache variables that is sent to the create case URL to create a case. Required variables are case.title and case.description.

For a Webhook - Case Management connector, specifies a REST API URL string to create a case in the third-party system.

The default model to use for requests, which varies by connector:

For a Webhook - Case Management connector, specifies a REST API URL string with an external service ID Mustache variable to get the case from the third-party system.

For a Webhook - Case Management connector, specifies a stringified JSON payload with Mustache variables that is sent to the update case URL to update a case. Required variables are case.title and case.description.

For a Webhook - Case Management connector, specifies the REST API URL to update the case by ID in the third-party system.

A configuration URL that varies by connector:

Sensitive configuration details, such as username, password, and keys, which are specific to the connector type.

For an Google Gemini connector, specifies the GCP service account credentials JSON file for authentication.

A client secret that varies by connector:

An email address that varies by connector:

A password secret that varies by connector:

For an xMatters connector with URL authentication, specifies the request URL for the Elastic Alerts trigger in xMatters with the API key included in the URL. It is used only when xpack.actions.preconfigured.<connector-id>.config.usesBasic is false.

A token secret that varies by connector:

A user name secret that varies by connector:

A URL that varies by connector:

Specifies the maximum number of alerts that a rule can generate each time detection checks run. Default: 1000.

Overrides the configs under xpack.alerting.rules.run for the rule type with the given ID. List the rule identifier and its settings in an array of objects. For example:

xpack.alerting.rules.run:
    timeout: '5m'
    ruleTypeOverrides:
        - id: '.index-threshold'
          timeout: '15m'

Overrides the configs under xpack.alerting.rules.run.actions for the connector type with the given ID. List the connector type identifier and its settings in an array of objects. For example:

xpack.alerting.rules.run:
    actions:
        max: 10
        connectorTypeOverrides:
            - id: '.server-log'
              max: 5