I’ve used JMeter for stress/load testing Mendix apps and it worked well.
My biggest tip: start by creating realistic test data (often mocked or generated) so your scenarios behave like production. Then, make sure you have REST endpoints (or other stable entry points) for the key user journeys you want to test, and map those endpoints directly to your test scenarios.
From there it’s mostly about configuring JMeter: thread groups (virtual users), ramp-up, think times, and realistic request pacing. JMeter also makes results easier to consume because it can generate HTML dashboards/reports. I usually export screenshots from those dashboards and share them with the PO, so we can discuss performance impact in a user-story context.