Computing dashboard data
Overview
The Portal employs different techniques to retrieve and process dashboard data from the connected Engines. Depending on their mode of activation, the techniques for computing dashboard data can be classified as automatic or manual:
Automatic computations
Computation after metric creation
Nightly computation
Real-time computation (live service data)
Manual computations
Launch computation of metric from the Finder
Launch computation of software metering metric from the Portal
In this article learn about the different types of computations, how to manually launch the computation of metrics from either the Finder or the Portal, and how to track the status of computations from the Portal. Learn also about the consequences of computing the values of a metric for dates in the past.
Automatic computations
Computation after metric creation
When you create and save a new metric in the Finder, it automatically launches its computation for the past day. A newly created metric has thus one day of history unless you clear the metric manually.
Nightly computation
The nightly computation is the main source of dashboard information for the Portal. It usually takes place during the night (hence its name), when the load on the Portal and on the Engines is lower because users seldom connect at night. At the scheduled time, the Portal starts collecting data related to the enabled services and metrics in all connected Engines and processes it for being displayed in the dashboards. This automatic computation is scheduled by default at 1 am local time, but this time is configurable.
The Portal determines the past day individually for each Engine in the following way:
The Portal gets the time of the last event stored in the Engine.
The past day is the whole day before the date of the last event in the Engine (always computed in the time zone of the Portal).
For Engines with low activity, this way of determining the past day might mean that the past day is two (or more) days ago. For instance, if an Engine has no event during the current day yet (when the Portal starts the nightly computation) but the last event happened the day before, the Portal considers the past day to be two days ago for that Engine.
Real-time computation
When displaying live data in service dashboards, the Portal continuously receives data from the Engines about the current status of the services. This automatic computation requires no special configuration.
Manual computations
Manual computations do not replace the nightly computation of the Portal, but are a complement to it. Manually computing a metric is specially useful when you create or modify a metric. Instead of waiting for the nightly computation to get the results for the metric, you can manually trigger its computation to see how it looks in the Portal.
Metric computation from the Finder
To trigger the computation of metrics from the Finder:
Right-click the name of a metric or of a folder holding metrics in the left-hand side accordion.
Select Compute from the context menu:
Choose For the last day, to compute or recompute the results of the metric (or metrics) for the day before.
A dialog warns you that the operation will clear the history of the metric (or metrics) for the last day. Click Yes to proceed.
A final dialog informs you that the computation of the metric (or metrics) will start shortly. Click OK.
Choose Over the maximum available period, to compute the metric (or metrics) for all the past days available in each connected Engine. This option is not available for count metrics that take into account all objects (active and inactive).
A dialog warns you that the operation will clear all the history of the selected metric. If you selected a folder, the dialog lists all the metrics in the selected folder, indicating those count metrics in the folder that take into account all objects, which are ignored for the computation. Click Yes to proceed.
A final dialog informs you that the computation of the metric (or metrics) will start shortly. Click OK.
Beware of the option to recompute a metric over the maximum available period. It really clears all the history of a metric. For instance, if you have three months of historical data in the Portal for a particular metric and two weeks of data available in the Engine, asking to recompute the metric over the full period will erase the three months of history and recreate only two weeks. Therefore, use this option with care and only when the modification of a metric completely invalidates previous results. Keep in mind as well that the recomputation of metrics for past days suffers from some limitations.
Software metering computation from the Portal
To compute software metering metrics that are already created:
Log in to the Portal as administrator.
Click the ADMINISTRATION item in the top menu.
Select Software metering metrics under CONTENT MANAGEMENT in the drop-down menu.
In the list of available metrics, tick the box in the first column for each metric that you want to compute. Alternatively, tick the box at the top of the list to select all available metrics at once.
Click the sprocket icon placed in the top right corner of the dashboard. A dialog informs you that the computation will be done for the last day.
Optional: Tick the box in the dialog to overwrite current data; that is, to clear and recompute as well those metrics that already have results for the last day.
Click OK to schedule the computation. Another dialog informs you that the computation is scheduled and that you can track it.
Click Done.
As any other metric, software metering metrics are also computed during the nightly computation of the Portal.
Tracking computations
To track the computations of metrics in the Portal:
Log in to the Portal as administrator.
Click the ADMINISTRATION item in the top menu.
Select Computations under CONTENT MANAGEMENT in the drop-down menu.
The Computations dashboard displays a table that includes both the manual and the automatic computations launched today by default. Filter the results by selecting other time frames and the user who initiated the computation from the two lists interleaved in the following sentence that you find at the top of the table:
Display computations for -time frame- created by -user.
As time frame, choose among:
today (default)
last week
last month
To filter by the user who launched the computation, choose among three options:
everybody (default): list all the computations, no matter who initiated them.
myself: list only the computations initiated by the current user.
Currently not working as intended, as the Nexthink Portal indirectly initiates all computations and is therefore displayed as owner of every computation.
The table shows the current status of each computation within the selected time frame, including (when applicable) its start time, end time, progress ratio, duration, and owner. The dashboard content is automatically refreshed every 10 seconds.
It is possible to stop a scheduled or running computation. Administrators can stop only those computations initiated by themselves. Central administrators, on the other hand, can stop computations initiated by anyone. If you stop a computation during its execution, the results for those dates that have been fully computed are saved and the rest are discarded.
The following table describes the possible computation statuses:
Scheduled
When a computation is scheduled, the operation has been submitted and put in the queue. There is no ongoing computation yet; therefore, there are no details about progress, time, or duration, but the computation can be cancelled.
Running
The computation is currently being performed. The start time and progress fields are available. The computation can be cancelled.
Completed (success)
All metrics were computed without error. Details about the computed metrics are available.
Completed (failure)
Errors occurred during the computation. Some metrics may have been computed correctly but at least one had an error. Details show the completed and failed computations of the metrics.
Cancelled
The computation has been manually cancelled. Details may show some metrics that have been completely computed before cancellation (with or without errors), while other metrics have not been computed.
Display the details of all finished computations (either completed or cancelled) by clicking the information button.
Computation details
To get details about a specific computation in the Computations dashboard:
Click the info button in the table, when the display mode is all computations.
Click any point in the line chart (or the info button to the right of a value), when the display mode is automatic computations.
The details dialog displays the list of metrics with their individual computation status. You can only see the details of those metrics that fall into your administration domain.
In the all computations mode, the table displays individual computations, including manual computations and separate attempts of nightly computations. Therefore, when you open the computation details from this mode, the date is fixed and there is no selection of attempts. On the other hand, when opening the details of a nightly computation from the automatic computations mode, it is possible to select a different date for convenience and see a different. In addition, ef multiple attempts were needed for the nightly computation of the selected date, you can also select the specific attempt. Thus, when opening the details dialog from the automatic computations mode, find these additional selectors:
Display details for -date: to select a particular date.
This computation required multiple executions. Display -attempt: to select a particular attempt.
In both modes, the dialog lets you choose the metrics for which you want to see the details, depending on the status of their computation:
Completed (successfully or not, but not skipped)
Skipped
Successful
Unsuccessful
For unsuccessfully computed metrics, click the cross icon to display the error that stopped the computation.
Widgets with no data
If the Portal has not computed data for a metric on a particular date, the widgets that display the value of the metric show it in different ways depending on their visual appearance:
In KPI or table widgets, the value of a metric with no data is represented with a dash (-) character.
In line charts, when hovering the mouse over the points where the metric was not computed display the message no data computed.
In the widgets of service dashboards, you find the same behavior when there is no data for the service. Similarly, in the overview dashboard of a service module, a dash character representing the status of a service indicates a lack of data for that particular service.
Note that you still may see live data in a service dashboard that received no data from nightly computations. For instance, the first day that you create a service, you may see live data for that service, but you will get no data for the service if you navigate to a previous date.
Computing metrics for dates in the past
There are a few cases where you may want to compute the value of metric for a date in the past:
You create a new metric and you want to compute its value for some days before its creation.
You modify a metric in such a way that it completely invalidates previously computed values. You may want to clear the history of the metric and recompute it for some days in the past.
The nightly computation of the Portal failed for some reason at a specific date and you want to have the value of a metric for that date.
You can either trigger the computation of a metric manually from the Finder or wait for a new nightly computation of the Portal. In the case of newly created metrics, only manually triggering the computation will do, because they automatically compute the last day at the time of creation, preventing thus the Portal from going further to the past. Remember that the nightly computation automatically recomputes the metrics up to the configured number of past days, if the data of the metrics are missing for some consecutive days including the last day. The nightly computation will not automatically fill in the gaps though if the failed computations are placed between successful computations.
Regardless of computing the metrics manually or automatically, you must be aware of the limitations of computing metrics for dates in the past.
Metrics may include in their computation values of objects (or conditions on these values) for which the Engine does not keep a historical record. Examples of these values may be the free disk space or the antivirus status of a device. During the nightly computation, the Portal takes the current value of those fields to compute the metrics for that day. In this way, the Portal always uses the most recent values to compute the metrics for the last day.
When computing the metrics for past dates, however, the Portal does not know the values of these fields in the past, so it still uses their most recent values. Since these fields usually do not change quickly over time, the computed values are often a good approximation, but if the value has a significant change (for instance, the antivirus real-time protection goes from on to off) the value of the metric may be wrong.
Therefore, beware of computing metrics for dates in the past when the metrics depend on values of fields for which the Engine does not keep a historical record.
RELATED TASK
RELATED REFERENCES
Last updated
Was this helpful?
