mirror of
https://github.com/elastic/kibana.git
synced 2025-04-24 01:38:56 -04:00
4c281a8995
* Update machine-learning.asciidoc * Update docs/apm/machine-learning.asciidoc Co-authored-by: Giorgos Bamparopoulos <georgios.bamparopoulos@elastic.co> Co-authored-by: Giorgos Bamparopoulos <georgios.bamparopoulos@elastic.co>
73 lines
3.1 KiB
Text
73 lines
3.1 KiB
Text
[role="xpack"]
|
|
[[machine-learning-integration]]
|
|
=== Machine learning integration
|
|
|
|
++++
|
|
<titleabbrev>Integrate with machine learning</titleabbrev>
|
|
++++
|
|
|
|
The Machine learning integration initiates a new job predefined to calculate anomaly scores on APM transaction durations.
|
|
With this integration, you can quickly pinpoint anomalous transactions and see the health of
|
|
any upstream and downstream services.
|
|
|
|
Machine learning jobs are created per environment, and are based on a service's average response time.
|
|
Because jobs are created at the environment level,
|
|
you can add new services to your existing environments without the need for additional machine learning jobs.
|
|
|
|
Results from machine learning jobs are shown in multiple places throughout the APM app:
|
|
|
|
* The **Services overview** provides a quick-glance view of the general health of all of your services.
|
|
+
|
|
[role="screenshot"]
|
|
image::apm/images/service-quick-health.png[Example view of anomaly scores on response times in the APM app]
|
|
|
|
* The transaction duration chart will show the expected bounds and add an annotation when the anomaly score is 75 or above.
|
|
+
|
|
[role="screenshot"]
|
|
image::apm/images/apm-ml-integration.png[Example view of anomaly scores on response times in the APM app]
|
|
|
|
* Service maps will display a color-coded anomaly indicator based on the detected anomaly score.
|
|
+
|
|
[role="screenshot"]
|
|
image::apm/images/apm-service-map-anomaly.png[Example view of anomaly scores on service maps in the APM app]
|
|
|
|
[float]
|
|
[[create-ml-integration]]
|
|
=== Enable anomaly detection
|
|
|
|
To enable machine learning anomaly detection:
|
|
|
|
. From the Services overview, Traces overview, or Service Map tab,
|
|
select **Anomaly detection**.
|
|
|
|
. Click **Create ML Job**.
|
|
|
|
. Machine learning jobs are created at the environment level.
|
|
Select all of the service environments that you want to enable anomaly detection in.
|
|
Anomalies will surface for all services and transaction types within the selected environments.
|
|
|
|
. Click **Create Jobs**.
|
|
|
|
That's it! After a few minutes, the job will begin calculating results;
|
|
it might take additional time for results to appear on your service maps.
|
|
Existing jobs can be managed in *Machine Learning jobs management*.
|
|
|
|
[float]
|
|
[[warning-ml-integration]]
|
|
=== Anomaly detection warning
|
|
|
|
To make machine learning as easy as possible to set up,
|
|
the APM app will warn you when filtered to an environment without a machine learning job.
|
|
|
|
[role="screenshot"]
|
|
image::apm/images/apm-anomaly-alert.png[Example view of anomaly alert in the APM app]
|
|
|
|
[float]
|
|
[[unkown-ml-integration]]
|
|
=== Unknown service health
|
|
|
|
After enabling anomaly detection, service health may display as "Unknown". There are three reasons why this can occur:
|
|
|
|
1. No machine learning job exists. See <<create-ml-integration>> to enable anomaly detection and create a machine learning job.
|
|
2. There is no machine learning data for the job. If you just created the machine learning job you'll need to wait a few minutes for data to be available. Alternatively, if the service or its enviroment are new, you'll need to wait for more trace data.
|
|
3. No "request" or "page-load" transaction type exists for this service; service health is only available for these transaction types.
|