The execution environment provided for the input includes includes the functions, macros, and global variables provided by the mito library. A single JSON object is provided as an input accessible through a state variable. state contains a response map field and may contain arbitrary other fields configured via the input’s state configuration. If the CEL program saves cursor states between executions of the program, the configured state.cursor value will be replaced by the saved cursor prior to execution.

On start the state will be something like this:

{
    "response": { ... },
    "cursor": { ... },
    ...
}

The websocket input creates a response field in the state map and attaches the websocket message to this field. All CEL programs written should act on this response field. Additional fields may be present at the root of the object and if the program tolerates it, the cursor value may be absent. Only the cursor is persisted over restarts, but all fields in state are retained between iterations of the processing loop except for the produced events array, see below.

If the cursor is present the program should process or filter out responses based on its value. If cursor is not present all responses should be processed as per the program’s logic.

After completion of a program’s execution it should return a single object with a structure looking like this:

{
    "events": [ 
        {...},
        ...
    ],
    "cursor": [ 
        {...},
        ...
    ]
}

Example configuration:

filebeat.inputs:
# Read and process simple websocket messages from a local websocket server
- type: websocket
  url: ws://localhost:443/v1/stream
  program: |
    bytes(state.response).decode_json().as(inner_body,{
      "events": {
        "message":  inner_body.encode_json(),
      }
    })

The Websocket input will log the complete state when logging at the DEBUG level before and after CEL evaluation. This will include any sensitive or secret information kept in the state object, and so DEBUG level logging should not be used in production when sensitive information is retained in the state object. See redact configuration parameters for settings to exclude sensitive fields from DEBUG logs.

The Websocket input supports authentication via Basic token authentication, Bearer token authentication and authentication via a custom auth config. Unlike REST inputs Basic Authentication contains a basic auth token, Bearer Authentication contains a bearer token and custom auth contains any combination of custom header and value. These token/key values are are added to the request headers and are not exposed to the state object. The custom auth configuration is useful for constructing requests that require custom headers and values for authentication. The basic and bearer token configurations will always use the Authorization header and prepend the token with Basic or Bearer respectively.

Example configurations with authentication:

filebeat.inputs:
- type: websocket
  auth.basic_token: "dXNlcjpwYXNzd29yZA=="
  url: wss://localhost:443/_stream

filebeat.inputs:
- type: websocket
  auth.bearer_token: "dXNlcjpwYXNzd29yZA=="
  url: wss://localhost:443/_stream

filebeat.inputs:
- type: websocket
  auth.custom:
    header: "x-api-key"
    value: "dXNlcjpwYXNzd29yZA=="
  url: wss://localhost:443/_stream

filebeat.inputs:
- type: websocket
  auth.custom:
    header: "Auth"
    value: "Bearer dXNlcjpwYXNzd29yZA=="
  url: wss://localhost:443/_stream

The websocket input keeps a runtime state between every message received. This state can be accessed by the CEL program and may contain arbitrary objects. The state must contain a response map and may contain any object the user wishes to store in it. All objects are stored at runtime, except cursor, which has values that are persisted between restarts.

The websocket input supports the following configuration options plus the Common options described later.

programedit

The CEL program that is executed on each message received. This field should ideally be present but if not the default program given below is used.

program: |
  bytes(state.response).decode_json().as(inner_body,{
    "events": {
      "message":  inner_body.encode_json(),
    }
  })

stateedit

state is an optional object that is passed to the CEL program on the first execution. It is available to the executing program as the state variable. Except for the state.cursor field, state does not persist over restarts.

state.cursoredit

The cursor is an object available as state.cursor where arbitrary values may be stored. Cursor state is kept between input restarts and updated after each event of a request has been published. When a cursor is used the CEL program must either create a cursor state for each event that is returned by the program, or a single cursor that reflects the cursor for completion of the full set of events.

filebeat.inputs:
# Read and process simple websocket messages from a local websocket server
- type: websocket
  url: ws://localhost:443/v1/stream
  program: |
    bytes(state.response).as(body, {
      "events": [body.decode_json().with({
        "last_requested_at": has(state.cursor) && has(state.cursor.last_requested_at) ?
          state.cursor.last_requested_at
        :
          now
      })],
      "cursor": {"last_requested_at": now}
    })

regexpedit

A set of named regular expressions that may be used during a CEL program’s execution using the regexp extension library. The syntax used for the regular expressions is RE2.

filebeat.inputs:
- type: websocket
  # Define two regular expressions, 'products' and 'solutions' for use during CEL program execution.
  regexp:
    products: '(?i)(Elasticsearch|Beats|Logstash|Kibana)'
    solutions: '(?i)(Search|Observability|Security)'

redactedit

During debug level logging, the state object and the resulting evaluation result are included in logs. This may result in leaking of secrets. In order to prevent this, fields may be redacted or deleted from the logged state. The redact configuration allows users to configure this field redaction behaviour. For safety reasons if the redact configuration is missing a warning is logged.

In the case of no-required redaction an empty redact.fields configuration should be used to silence the logged warning.

- type: websocket
  redact:
    fields: ~

As an example, if a user-constructed Basic Authentication request is used in a CEL program the password can be redacted like so

filebeat.inputs:
- type: websocket
  url: ws://localhost:443/_stream
  state:
    user: user@domain.tld
    password: P@$$W0₹D
  redact:
    fields:
      - password
    delete: true

Note that fields under the auth configuration hierarchy are not exposed to the state and so do not need to be redacted. For this reason it is preferable to use these for authentication over the request construction shown above where possible.

redact.fieldsedit

This specifies fields in the state to be redacted prior to debug logging. Fields listed in this array will be either replaced with a * or deleted entirely from messages sent to debug logs.

redact.deleteedit

This specifies whether fields should be replaced with a * or deleted entirely from messages sent to debug logs. If delete is true, fields will be deleted rather than replaced.

Metricsedit

This input exposes metrics under the HTTP monitoring endpoint. These metrics are exposed under the /inputs path. They can be used to observe the activity of the input.

A stand-alone CEL environment that implements the majority of the websocket input’s Comment Expression Language functionality is available in the Elastic Mito repository. This tool may be used to help develop CEL programs to be used by the input. Installation is available from source by running go install github.com/elastic/mito/cmd/mito@latest and requires a Go toolchain.

The following configuration options are supported by all inputs.

enablededit

Use the enabled option to enable and disable inputs. By default, enabled is set to true.

tagsedit

A list of tags that Filebeat includes in the tags field of each published event. Tags make it easy to select specific events in Kibana or apply conditional filtering in Logstash. These tags will be appended to the list of tags specified in the general configuration.

Example:

filebeat.inputs:
- type: websocket
  . . .
  tags: ["json"]

fieldsedit

Optional fields that you can specify to add additional information to the output. For example, you might add fields that you can use for filtering log data. Fields can be scalar values, arrays, dictionaries, or any nested combination of these. By default, the fields that you specify here will be grouped under a fields sub-dictionary in the output document. To store the custom fields as top-level fields, set the fields_under_root option to true. If a duplicate field is declared in the general configuration, then its value will be overwritten by the value declared here.

filebeat.inputs:
- type: websocket
  . . .
  fields:
    app_id: query_engine_12

fields_under_rootedit

If this option is set to true, the custom fields are stored as top-level fields in the output document instead of being grouped under a fields sub-dictionary. If the custom field names conflict with other field names added by Filebeat, then the custom fields overwrite the other fields.

processorsedit

A list of processors to apply to the input data.

See Processors for information about specifying processors in your config.

pipelineedit

The ingest pipeline ID to set for the events generated by this input.

keep_nulledit

If this option is set to true, fields with null values will be published in the output document. By default, keep_null is set to false.

indexedit

If present, this formatted string overrides the index for events from this input (for elasticsearch outputs), or sets the raw_index field of the event’s metadata (for other outputs). This string can only refer to the agent name and version and the event timestamp; for access to dynamic fields, use output.elasticsearch.index or a processor.

Example value: "%{[agent.name]}-myindex-%{+yyyy.MM.dd}" might expand to "filebeat-myindex-2019.11.01".

publisher_pipeline.disable_hostedit

By default, all events contain host.name. This option can be set to true to disable the addition of this field to all events. The default value is false.