See at adding some useful metrics/traces in one router * [ ] understand why traces appear multiple times  * [ ] Make meaningful plots (timing distribution per endpoint)