opc results commands such as opc results pipelinerun list and opc results taskrun list accept --limit option, which can be used to limit the number of results returned. It currently can be between 5 and 1000, with the default of 50.

If one wants to ensure that "all" results are returned, it is possible to use the upper boundary --limit 1000. Is that recommended form the server point of view, or can that lead to a too much load per request? Would it be preferred to have client request the data in smaller batches?

For smaller batches, or cases when there are more than 1000 results, a proper pagination support would be required. I.e. not only set limit, but also some sort of "start at" indicator, allowing user to request e.g. results 1-100, 101-200, 201-300, ...