Hey Daniel,
It is actually not us being not able to keep up, it is Kubernetes not being able to keep up with itself (we used to use the official Python client to query the Kubernetes API, which always lags behind it).
That’s why we changed the approach for the new Kubernetes monitoring. There is a new ruleset called Kubernetes in 2.1.0. The old one is now called Kubernetes (deprecated). You need to configure the new ruleset and get rid off the old deprecated one.
Follow the instructions on checkmk_kube_agent/deploy/charts/checkmk at 1.0.0 · tribe29/checkmk_kube_agent · GitHub
The new monitoring has been tested on K8s 1.21 to 1.23. This is also the official support scope of the new monitoring (latest 3 versions).