We got a user asking about collector crash looping on clusters that using a RLIMIT_MEMLOCK of 8MB: https://redhat-internal.slack.com/archives/C01R0E7CVMX/p1749502664447509

This issue is most likely due to collector using the available physical memory of the system to determine the sizes of the bpf buffers it will allocate and ignoring this RLIMIT_MEMLOCK value.

In order to make it easier for customers deploying collector in environments with this limit set, we should make collector aware of it and adjust the size of buffers used.