Optimizing your Views
You can create highly customizable Views of your data in the Celonis Platform, giving you the insights you need into your business. However, there is a 3MB View size limit to ensure performance and stability, meaning that optimizing your Views is essential.
To optimize your Views and ensure they run smoothly, we recommend the following best practice:
Create focused packages and Knowledge Models: Rather than creating all Views and Knowledge Models in one package, try to focus your packages around your use cases. By having use case specific packages, you require less data to be loaded when opening your Views. We also recommend that a package contains no more than four Knowledge Models.
Limit your use of Linked Views: While linked Views allow you to display a greater volume of content, that data also needs to be loaded into your View each time. The loading time is also increased when embedded Views are then nested into other embedded Views.
For more information about linked Views, see: Linked Views.
Use buttons and links: Rather than using multiple tabs within your View, consider using buttons and links to direct your users to related Views instead. This approach reduces the load on individual Views, spreading out the data across more manageable Views.
For more information about buttons and links, see: Configuring a button.
Use View Modules: View modules are a package asset type that allow you to re-use configured View components across multiple Views within a Studio package. Instead of configuring each View from scratch, you can save time by centrally managing and then embedding a module to your View. And by editing a View module, all embedded instances of that module are automatically updated too.
To learn about View Modules, see: View modules.
Consider the volume of data needed: In general, the more data points used and the wider the timescale, the longer this data will take to load. As such, you should consider creating Views or setting your filters based on more defined time periods and use cases. As an example, rather than filtering a chart based on all data from the last year, use smaller time periods.
Optimize your charts: To ensure your charts display efficiently, we recommend limiting the number of data points to two thousand, taking advantage of Process Query Language (PQL) calculations (such as PU_Count), and defining both your x and y axis for each chart.
For more information, see: Charts.
Simplify your PQL and variable statements: When writing PQL queries and variable statements, we recommend keeping them targeted and using functions and short hands whenever possible. This will reduce the volume of information sent, improving the performance of your views.
We also recommend the following PQL focused best practice:
Try to aggregate on small tables first and use PU functions.
Apply your aggregations and FILTER_TO_NULL as late as possible.
Use process operators as opposed to non-process operators.
Use REMAP_VALUES instead of CASE WHEN.
Related links
Performance Optimization in PQL Academy course