Data retention

Data retention in the Portal

Nexthink is able to keep historical data for several years in the database of the Portal. The Portal consolidates the data collected from the Engines and keeps them in permanent storage. The Portal does not store all the individual events, but the results of widget computation.

History

Number of devices

Several years

(view by periods of 2 years max)

Up to 150 000

(consult for bigger setups)

Keeping historical details of count metrics

In the case of count metrics, the Portal stores the total number of objects satisfying a particular set of conditions. The Portal keeps these numbers of counted objects as regular historical data. In addition, for every count metric, the Portal stores the list of objects that contributed to the metric for the current day, week, month, and quarter. The list of objects that contributed to a count metric, along with their selected set of display fields, are known as the details of the metric. To see the details of a count metric, open a dashboard with KPI or table widgets representing the values of that count metric in the Portal, hover the mouse cursor over a particular value and select Show details.

When additional disk space is allocated, the Portal can store the details of count metrics not only for the current period (aggregated object lists for the current day, week, month, and quarter), but also for previous days, weeks, months, and quarters. To keep historical details of count metrics, make sure that you reserve some disk space for that purpose in the Portal appliance:

  1. Log in to the Web Console of the Appliance that hosts the Portal as admin. Use your web browser:https://:99

  2. Click the Portal tab at the top of the window.

  3. Select the General section from the left-hand side menu.

  4. Under Parameters, choose the quantity of disk space that you want to dedicate to the storage of history details for count metrics from the list labeled Disk space allocated for historical data.

The number of days of stored historical details depend on the amount of disk space reserved, the number of enabled count metrics, the number of display fields included in each metric, and the actual metric data collected each day. When the disk space dedicated to store the details of count metrics is exhausted, the Portal launches a cleanup process that deletes one or a few full days of historical details, starting from the oldest day in the saved history.

Find orientative figures for setting the disk space for historical details in the hardware requirements of the Portal.

Data retention in the Engine

In its turn, the Engine stores real-time data with a greater level of detail than the Portal. All events are kept in volatile memory until the configured maximum is reached. You can browse all the data in the Engine down to the event level with the help of the Finder.

We present in this section several tables that contain an estimation of the number of weeks that correspond to a maximum number of events stored the Engine. Once the maximum is reached, new events replace the oldest ones, which have already been consolidated in the Portal before being dismissed.

Estimations are calculated on the following basis:

  • Working day of 8 hours.

  • Week of 5 working days.

  • Practically no activity outside working hours.

Please note that the amount of data required for mobile devices is negligible.

Without the Web and Cloud Feature

If the Web and Cloud feature is not enabled, the data retention periods are expressed as follow (in weeks):

200M events

100M events

50M events

10 000 devices

4-6 weeks

2-3 weeks

1-2 weeks

5 000 devices

8-10 weeks

4-5 weeks

2-3 weeks

With Web and Cloud Feature but for Services Only

If Web and Cloud has been exclusively enabled for defined Services, the following data retention periods are expected (in weeks):

200M events

100M events

50M events

10 000 devices

4 weeks

2 weeks

1 week

5 000 devices

8 weeks

4 weeks

2 weeks

With Full Web and Cloud Feature

If the Web and Cloud feature is fully enabled, the following data retention periods are expected (in weeks):

200M events

100M events

50M events

10 000 devices

2-4 weeks

1-2 weeks

1 week

5 000 devices

4-6 weeks

2-3 weeks

1-2 weeks

Increasing data retention in the Engine

If you reckon that your current data retention period is rather short, increase it by trading off detailed information for more capacity. Gain up to 50% more history by setting a more aggressive aggregation policy in the Engine.


RELATED TASK

Last updated