Get categories API

edit

Retrieves anomaly detection job results for one or more categories.

Request

edit

GET _ml/anomaly_detectors/<job_id>/results/categories

GET _ml/anomaly_detectors/<job_id>/results/categories/<category_id>

Prerequisites

edit
  • If the Elasticsearch security features are enabled, you must have monitor_ml, monitor, manage_ml, or manage cluster privileges to use this API. You also need read index privilege on the index that stores the results. The machine_learning_admin and machine_learning_user roles provide these privileges. See Security privileges, Built-in roles, and Machine learning security privileges.

Description

edit

When categorization_field_name is specified in the job configuration, it is possible to view the definitions of the resulting categories. A category definition describes the common terms matched and contains examples of matched values.

The anomaly results from a categorization analysis are available as bucket, influencer, and record results. For example, the results might indicate that at 16:45 there was an unusual count of log message category 11. You can then examine the description and examples of that category. For more information, see Categorizing log messages.

Path parameters

edit
<category_id>
(Optional, long) Identifier for the category, which is unique in the job. If you specify neither the category ID nor the partition_field_value, the API returns information about all categories. If you specify only the partition_field_value, it returns information about all categories for the specified partition.
<job_id>
(Required, string) Identifier for the anomaly detection job.

Request body

edit
page.from
(Optional, integer) Skips the specified number of categories.
page.size
(Optional, integer) Specifies the maximum number of categories to obtain.
partition_field_value
(Optional, string) Only return categories for the specified partition.

Response body

edit

The API returns an array of category objects, which have the following properties:

category_id
(unsigned integer) A unique identifier for the category. category_id is unique at the job level, even when per-partition categorization is enabled.
examples
(array) A list of examples of actual values that matched the category.
grok_pattern
[preview] This functionality is in technical preview and may be changed or removed in a future release. Elastic will work to fix any issues, but features in technical preview are not subject to the support SLA of official GA features. (string) A Grok pattern that could be used in Logstash or an ingest pipeline to extract fields from messages that match the category. This field is experimental and may be changed or removed in a future release. The Grok patterns that are found are not optimal, but are often a good starting point for manual tweaking.
job_id
(string) Identifier for the anomaly detection job.
max_matching_length
(unsigned integer) The maximum length of the fields that matched the category. The value is increased by 10% to enable matching for similar fields that have not been analyzed.
partition_field_name
(string) If per-partition categorization is enabled, this property identifies the field used to segment the categorization. It is not present when per-partition categorization is disabled.
partition_field_value
(string) If per-partition categorization is enabled, this property identifies the value of the partition_field_name for the category. It is not present when per-partition categorization is disabled.
regex
(string) A regular expression that is used to search for values that match the category.
terms
(string) A space separated list of the common tokens that are matched in values of the category.
num_matches
(long) The number of messages that have been matched by this category. This is only guaranteed to have the latest accurate count after a job _flush or _close
preferred_to_categories
(list) A list of category_id entries that this current category encompasses. Any new message that is processed by the categorizer will match against this category and not any of the categories in this list. This is only guaranteed to have the latest accurate list of categories after a job _flush or _close

Examples

edit
GET _ml/anomaly_detectors/esxi_log/results/categories
{
  "page":{
    "size": 1
  }
}

In this example, the API returns the following information:

{
  "count": 11,
  "categories": [
    {
      "job_id" : "esxi_log",
      "category_id" : 1,
      "terms" : "Vpxa verbose vpxavpxaInvtVm opID VpxaInvtVmChangeListener Guest DiskInfo Changed",
      "regex" : ".*?Vpxa.+?verbose.+?vpxavpxaInvtVm.+?opID.+?VpxaInvtVmChangeListener.+?Guest.+?DiskInfo.+?Changed.*",
      "max_matching_length": 154,
      "examples" : [
        "Oct 19 17:04:44 esxi1.acme.com Vpxa: [3CB3FB90 verbose 'vpxavpxaInvtVm' opID=WFU-33d82c31] [VpxaInvtVmChangeListener] Guest DiskInfo Changed",
        "Oct 19 17:04:45 esxi2.acme.com Vpxa: [3CA66B90 verbose 'vpxavpxaInvtVm' opID=WFU-33927856] [VpxaInvtVmChangeListener] Guest DiskInfo Changed",
        "Oct 19 17:04:51 esxi1.acme.com Vpxa: [FFDBAB90 verbose 'vpxavpxaInvtVm' opID=WFU-25e0d447] [VpxaInvtVmChangeListener] Guest DiskInfo Changed",
        "Oct 19 17:04:58 esxi2.acme.com Vpxa: [FFDDBB90 verbose 'vpxavpxaInvtVm' opID=WFU-bbff0134] [VpxaInvtVmChangeListener] Guest DiskInfo Changed"
      ],
      "grok_pattern" : ".*?%{SYSLOGTIMESTAMP:timestamp}.+?Vpxa.+?%{BASE16NUM:field}.+?verbose.+?vpxavpxaInvtVm.+?opID.+?VpxaInvtVmChangeListener.+?Guest.+?DiskInfo.+?Changed.*"
    }
  ]
}