Abstract Tightness against media is a frequent requirement for technical components. Despite various standardized test procedures, failure regularly occurs during use. Often, no clear cause for failure can be determined afterwards. In this article, a new test setup is presented and applied, which allows an in-situ measurement during a mechanical load. A flow meter with a measuring range of 0.02 mL/min to 5 mL/min is used for this purpose. This makes it possible to determine leakage rates with time resolution and thus to identify the moment of failure or the causal failure load. This new method was applied directly to different aluminum inserts with a polyamide 66 (PA66) overmold. It was shown that no increase in leakage occurs until a maximum force is reached, even with multiple loads. This maximum force depends only on the pretreatment of the inserts and can be determined in a simple pull-out test independently of the test setup used here and can therefore be used for the design of assemblies. In the test, a maximum force of 100 N was achieved for untreated inserts, 140 N for adhesion promoter-coated parts and 600 N for etched inserts with a contact area of 48 mm2. With this results, a new link between adhesion and tightness can be shown, which of course is only valid for initial tight parts.