This is not the same as RIC-30 (which is about latency measurements)
We dropped RIC-126 from R4. I.e. no collection via RMR. The reporting to Prometheus via Prometheus clients libraries would already work, but this path we did not want to take for E2T. What's needed in RIC-178 is simply the infrastructure to count number of ingress and egress messages as a cumulative monotonically increasing counter. Maintained per gNB. How we report it is not that important for now, as eventually we will connect it to a statistics collection system (once that supports RMR reporting).
As suggested in e-mails, since we don't focus on scalability (in terms of numbers of RAN nodes), let's simply write a log entry and report the current counter value every X minutes. Maybe make X a config parameter (in 100ms steps) of E2T which defaults to 60000 milliseconds.