2019-08-14 02:26:03 -04:00
|
|
|
[role="xpack"]
|
|
|
|
[testenv="platinum"]
|
|
|
|
[[estimate-memory-usage-dfanalytics]]
|
|
|
|
=== Estimate memory usage API
|
|
|
|
|
|
|
|
[subs="attributes"]
|
|
|
|
++++
|
|
|
|
<titleabbrev>Estimate memory usage for {dfanalytics-jobs}</titleabbrev>
|
|
|
|
++++
|
|
|
|
|
|
|
|
Estimates memory usage for the given {dataframe-analytics-config}.
|
|
|
|
|
|
|
|
experimental[]
|
|
|
|
|
|
|
|
[[ml-estimate-memory-usage-dfanalytics-request]]
|
|
|
|
==== {api-request-title}
|
|
|
|
|
|
|
|
`POST _ml/data_frame/analytics/_estimate_memory_usage`
|
|
|
|
|
|
|
|
[[ml-estimate-memory-usage-dfanalytics-prereq]]
|
|
|
|
==== {api-prereq-title}
|
|
|
|
|
|
|
|
* You must have `monitor_ml` privilege to use this API. For more
|
2019-10-07 18:23:19 -04:00
|
|
|
information, see <<security-privileges>> and <<built-in-roles>>.
|
2019-08-14 02:26:03 -04:00
|
|
|
|
|
|
|
[[ml-estimate-memory-usage-dfanalytics-desc]]
|
|
|
|
==== {api-description-title}
|
|
|
|
|
|
|
|
This API estimates memory usage for the given {dataframe-analytics-config} before the {dfanalytics-job} is even created.
|
|
|
|
|
|
|
|
Serves as an advice on how to set `model_memory_limit` when creating {dfanalytics-job}.
|
|
|
|
|
|
|
|
[[ml-estimate-memory-usage-dfanalytics-request-body]]
|
|
|
|
==== {api-request-body-title}
|
|
|
|
|
|
|
|
`data_frame_analytics_config`::
|
|
|
|
(Required, object) Intended configuration of {dfanalytics-job}. For more information, see
|
|
|
|
<<ml-dfanalytics-resources>>.
|
|
|
|
Note that `id` and `dest` don't need to be provided in the context of this API.
|
|
|
|
|
|
|
|
[[ml-estimate-memory-usage-dfanalytics-results]]
|
|
|
|
==== {api-response-body-title}
|
|
|
|
|
2019-08-21 08:27:36 -04:00
|
|
|
`expected_memory_without_disk`::
|
2019-08-14 02:26:03 -04:00
|
|
|
(string) Estimated memory usage under the assumption that the whole {dfanalytics} should happen in memory
|
|
|
|
(i.e. without overflowing to disk).
|
|
|
|
|
2019-08-21 08:27:36 -04:00
|
|
|
`expected_memory_with_disk`::
|
2019-08-14 02:26:03 -04:00
|
|
|
(string) Estimated memory usage under the assumption that overflowing to disk is allowed during {dfanalytics}.
|
2019-08-21 08:27:36 -04:00
|
|
|
`expected_memory_with_disk` is usually smaller than `expected_memory_without_disk` as using disk allows to
|
|
|
|
limit the main memory needed to perform {dfanalytics}.
|
2019-08-14 02:26:03 -04:00
|
|
|
|
|
|
|
[[ml-estimate-memory-usage-dfanalytics-example]]
|
|
|
|
==== {api-examples-title}
|
|
|
|
|
2019-09-09 12:35:50 -04:00
|
|
|
[source,console]
|
2019-08-14 02:26:03 -04:00
|
|
|
--------------------------------------------------
|
|
|
|
POST _ml/data_frame/analytics/_estimate_memory_usage
|
|
|
|
{
|
|
|
|
"data_frame_analytics_config": {
|
|
|
|
"source": {
|
|
|
|
"index": "logdata"
|
|
|
|
},
|
|
|
|
"analysis": {
|
|
|
|
"outlier_detection": {}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
--------------------------------------------------
|
|
|
|
// TEST[skip:TBD]
|
|
|
|
|
|
|
|
The API returns the following results:
|
|
|
|
|
2019-09-06 09:22:08 -04:00
|
|
|
[source,console-result]
|
2019-08-14 02:26:03 -04:00
|
|
|
----
|
|
|
|
{
|
2019-08-21 08:27:36 -04:00
|
|
|
"expected_memory_without_disk": "128MB",
|
|
|
|
"expected_memory_with_disk": "32MB"
|
2019-08-14 02:26:03 -04:00
|
|
|
}
|
|
|
|
----
|