I will try to keep a long story short. It all started with the following patron with routeros v6.7

In Green the uptime and the other colors are the load of each core.
(notice at first I only monitored the first core. I add mid-march the other cores after investigation)
Conclusion, the system gets busier when the uptimes gets higher.
Let’s look at the traffic patrons during that period.


Not the same patron, doesn’t grow and goes low again like the cpu load.
If I magnify the memory use vs uptime during the same period I get this

Same like cpu load. I know it’s not a lot of memory difference, but same patron.
To solve the issue I tried some more v6 version:

Strange things, I want to understand why the cpu load rises and the traffic doesn’t. Downgrade to v5.26 show the old trends I always say.

Long term trend view… (please ignore the DDOS period)
The router is running about 8 bgp peers, 2 ospf peers, 300k bgp routes, 50 ospf routes, no conntrack, x86, 20 forward firewall rules.
I can’t help that I think something is going wrong. I understand that v5 and v6 report a different value for the same traffic but it should be linked to the load / use and should not build up when the uptime becomes more. The problem is not really an issue since the router can handle the load. When v6.13 is available i will test it again, for now i will keep running v5.26.
I will keep you all informed. Tip / suggestions are welcome about this problem.