I work for a company where the amount of users doing actions on the application is very minimal. So doing increasing users over time, does not replicate real world behaviour. The issues that cause performance related problems are more related to api calls that take x amount of time, if the databases of the clients are big. You see then that requests take longer in the metric overview that K6 provides.
I leave some room for the community answers and other k6-sers, but here are my thoughts.
I’d say, as usual, the answer is based on your needs. For example, what are you trying to figure out when performing testing?
Suppose it’s the point when your systems start experiencing problems. In that case, you probably need to have a workflow that produces the data for the test system and monitors how the system behaves, and when you start to see that the systems started to respond slowly, remember that number and start figuring out what to do with that.
As you mentioned, you already know that the issue is the extensive database, then you probably should think about how to resolve the issues. And k6 here will be more the tool that could validate whether actions taken are good, like whether we see improvements in the metrics and so on.
