A very interesting topic. Capella provide several ways to integrate non-functional (incl. performance) concerns to the system architecture model. The choice of a particular way will strongly depend on the number of requirements, their nature and of course your available tools. I will only list some of them, but I may have forgotten some others.
First, Capella allows you to attach properties (Property Values) to any element in your model. You can use them to define non-functional properties. Check these posts for more information:
In a few weeks a new tool that will ease the definition of properties (like UML profiles) will be freely available
You may also want to check the
Capella Requirements add-on. Besides of importing requirements in ReqIF format, it will allow you to define types of requirements, types of relations between requirements and with model elements, and to embed these requirements in Capella diagrams, like in the image below:
However, keep in mind that this add-on won’t replace a dedicated Requirements Management tool that provides many more features.
Speaking specifically about performance requirements, you may associate execution times to Functions and Functional Exchanges, then check if you Functional Chains’ max. execution time is ensured. Check the Basic Performance Viewpoint (
Last but not least, some non-functional requirements may be expressed at System Analysis level as dependencies between the System and the Actors. This may be the case for your “external loads”, which could be modeled as an Interface, or even as a dependency between a Function to be ensured by the system and a Function ensured by an Actor.
Hope it helps.