High time usage for network exec

So I built (at pandorabox server, public mine) bit over 3000 cable, 60 generator, ~25 consumer MV network.
About 100kEU MV hydro tower.
Few supply converters, maybe 7 - 8.
Some MV tool workshops, about 12 maybe.
2 - 4 MV alloy furnaces.
No battery boxes.

High cable count is because machines are scattered over a large area forming small clusters each with additional HV network drawing power from main MV network through supply converters.

Network generally reports about 2 - 4ms time usage but sometime for few moments a lot more, around 80 - 600ms.
High values for lag metrics seems to happen when network gets activated after being away from area for some time.

Find out what causes this and if it can be fixed somehow.

I suspect this but it can be something else too:
https://github.com/mt-mods/technic/blob/62d523e368c38d31a48e37084542acc2ebc51eb1/technic/machines/network.lua#L454-L458

Start in technic/machines/network.lua at lines 454-458 and reproduce network activation after returning to the area while watching the reported time usage. Trace the timing spike to its cause and verify that the high usage is resolved; the issue does not name a test.