(DI-1904) BW System Health

Run times and errors in process chain steps

The visualization displays run time and error statistics for process chain steps, which were running on a system in the selected time range. This analysis enables you to check all active process chains at once on a particular system.
You have a possibility of two view perspectives, by selecting "day" or "week" in distribution in. You may filter results by the displayed characteristics.
Basic view displays results in four levels:

Result KPI types

You may select in Display a sum or an average (average is possible only for values > 0) of the KPIs run times and number of errors.

Distribution

Results can be distributed and aggregated into a day view or a week view.

In the day view the results are displayed in hours and by the week view in days.

Run time value calculation

The value of a run time depends on the chosen sampling frequency, distribution and the selected time range. Value of a run time displays, how many frequency intervals of a process chain step were running in the selected time range.

Let's show it on an example:

A process chain step was running from 10:45 to 12:50 and has the following settings:

By distributing this process chain step in a day view we obtain 3 time intervals which intersect with the day view distribution .

With the change of the sampling frequency you can obtain more or less detailed results.

Query runtime

The visualizations show query average runtimes per some analyzed period. There are two charts visible. Left side chart displays query groups based on average runtime i.e. how many queries in system belong to each performance group during analyzed period. Right side chart displays scatter chart plot for each individual query with average runtime and number of executions. The user can filter data and also switch to the table view for more detailed information where data can be downloaded into a CSV file for further processing.

_{Query runtime analysis charts}

User can click (drill down) in the left column chart in order to get more information about query execution. 

_{Query execution details table}

Here it is also possible to filter or download data. User can also click (drill down) on the row in order to get more information about query execution as time when query was executed, min., max. and average query runtime and also selection string.

To this analysis it is also possible to go directly from the scatter plot in the Query runtime analysis (by drilling down on the data in the scatter plot).

_{Query execution details per filtering InfoObjects combinations}

Here it is also possible to filter or download data or switch to table view and download data to CSV file for further processing.

Query runtime by categories

The visualizations show query average runtimes according to query processing categories per some analyzed period. Displayed categories are BI Suite, Data Manager, Olap - Cache, Olap – Olap, Olap – Authorizations and Olap – Input Help (for more info). 
Top part of analysis display average time spent in processing by each category for whole system. While bottom chart displays average processing times in categories for individual queries. 

_{Query runtime by categories analysis chart}

The user can filter data and also switch to the table view for more detailed information where data can be downloaded into a CSV file for further processing.

The worst runtime queries

The visualization shows the worst and also best query executions runtimes per some analysed period per day. For each day maximum of 10 worst query executions are captured. 

_{Query worst runtime analysis chart}

The user can filter data and also switch to the table view for more detailed information where data can be downloaded into a CSV file for further processing.
It is possible to click on the data in the plot area or in the table and the user will be redirected to the exact selection of query for corresponding worst execution, where data can be filtered and downloaded into the CSV.

_{Selection filters of worst query execution}