Now that this testnet is winding down I’ve taken the graphs generated from @Josh 's NTracking system. NTracking
Archive.zip (7.1 MB)
Here are some highlights
TCP Connections
Something happened on 21/10 lasting for more than a day. More than one node had a big drop in tcp connections.
Bubble Rewards
There’s the big boy. But there were a couple of other nodes that were doing well.
Rewards Balance
My super earner node with 160 Toy Town Tokens dwarfed all other nodes apart from one that had 43.
You have to switch off the top 3 to be able to see the growth of the others.
All we can say for sure is that rewards over time for a node and rewards over record count are unpredictable. Something that is apparently doing poorly can get a rocket up it. This may all be due to the smallness of the network.
Record Count
The record count was actually quite even in the end.
Memory Usage
There is still some kind of memory leak but it’s a lot less than before. Those big peaks for a few nodes tie in with when there were a lot of GETs. The values did decrease but not completely.
I started 40 nodes on a AWS T4g.medium Instance with 4GB RAM. I was kicking myself for not being bolder and starting 50 on it when it was clear the memory leak issue was a lot less. However, but the end there was only 500MB available so I’d have had some nodes killed.