Added the ability to create and display forecasts for machine learning jobs in Kibana.
Simplified job creation for certain recognizable data sources, such as
Filebeat access logs from Nginx and Apache HTTP servers.
Added a Data Visualizer, which enables you to learn more about the
characteristics of your data and identify fields for machine learning analysis. For more
information, see Creating Machine Learning Jobs.
Added bucket span to the chart labels in the Single Metric Viewer.
Added support for job groups when managing jobs in Kibana.
Added the ability to show or hide the Anomaly Explorer charts by selecting a
checkbox. The choice is saved in the URL, so it persists if you re-open the
view from a bookmarked URL.
Added a new job creation wizard for population jobs, which detect activity
that is unusual compared to the behavior of the population. For more information,
see Performing Population Analysis.
Added the ability to control the maximum number of rows in the Anomaly
Explorer view by swimlane, with options of 5, 10, 25, or 50. Previously,
it was fixed to a maximum of 10 rows.
Changed the job creation wizards such that you select an index pattern or
saved search before you select the type of job.
Added licensing checks to the job creation features in Kibana. If the license
is expired, for example, the create job, edit job, clone job, and start datafeed
buttons are disabled.
Added the ability to set the model memory limit in job creation wizards. For
more information about the model_memory_limit property, see
Job Resource Analysis Limits.
Added a Close Job button that enables you to force the closure of a job
when it fails. The button is disabled if the job is running, closed, or does not
have a datafeed.
Added the ability to display jobs with multiple detectors in the Single Metric
Viewer when the model_plot_config property is enabled. Previously, this
functionality was only possible for jobs with a single detector.
Added Cross Cluster Search (CCS) to efficiently retrieve monitoring data from
local and remote clusters.
Improved handling of health check errors. If there is a problem making a query,
for example, an error is returned and extra information is provided in the logs.
As part of this change, the monitoring plugin for Kibana will never cause Kibana
to go red.
Improved logging when X-Pack is not installed on the monitoring cluster.
Saved table interaction states such that if you leave the Kibana page and
subsequently return, your filters persist.
Added support for a HTTP Proxy to be configured when using Chromium as the
browser. For more information, see Chromium Settings.
Added headless build of Chromium that can be used instead of PhantomJS for
capturing Reporting screenshots. To enable Chromium, set the following setting
in your kibana.yml file: xpack.reporting.capture.browser.type: chromium. For
now PhantomJS will remain the default. In the future, however, Chromium will
become the default.
Added an option for PDF reports that preserves the existing layout and size of
the Visualization or Dashboard.
Added manage_index_templates cluster privileges to the kibana_system role,
which enables the Kibana system user to create and manage the index template for
the .kibana index. For more information, see
Security Privileges and
Updated the dates and times on the machine learning pages in Kibana, such that they use the
time zone information that is set in the dateFormat.tz property in the Kibana
Management Advanced Settings. Previously, dates and times were displayed in the
time zone of the web browser.
Added support for keyword and ip fields in multi-metric and population
job creation wizards.
Shortened custom urls to remove unnecessary information.
Added warning messages and disabled job creation when the index pattern does
not contain time series data.
Fixed situation where the watch generated by the job creation wizard was
generating an e-mail that contained incorrect time information.
Fixed memory leak in the overall swimlane for multi-metric and population
Fixed the View examples link that is provided for categorization anomalies.
In particular, fixed the link for datafeeds that use wildcards in their index
patterns and jobs that have nested fields in the categorization_field_name
Ensured that the aggregation interval that is used in the Single Metric charts
is a multiple of the job’s bucket span.
Improved the X-axis tick labels for machine learning charts and swimlanes, such that they
take into account the available chart width or screen space.
Added the median, high_median, and low_median functions to the single- and
multi-metric job creation wizards in Kibana.
Applied severity threshold settings to the Anomaly Explorer charts, such that
individual anomalies that do not meet the threshold are hidden.
Updated the dates and times on the X-Pack monitoring pages in Kibana, such that they
use the time zone information that is set in the dateFormat.tz property in the
Kibana Management Advanced Settings. Previously, dates and times were displayed
in the time zone of the web browser.
Fixed ability to show or hide the legend in monitoring charts.
Fixed internal server errors in Watcher that were caused by undefined job