From iftop and tcptrack -i eth0 I can see over 140 connections to[2a0c:dd40:1:b:0:0:0:42]:443, which is the IP of snowflake-01.torproject.net, and network speed to this IP is usually very high (about 500KB/s). So I assume Snowflake contributes to this statistics a lot.
You can see in the first days, traffic is not only big but increasing, that’s because Snowflake sometimes exits itself for a while and I’m deploying Systemd “restart” policy, to run it 24 hours per day.
However, don’t you think the traffic is too big? Running Snowflake 24 hours a days consumes about 100 GB traffic, though Snowflake itself says it’s just no more than 10GB in logs.
Thank you for running a Snowflake proxy. The high bandwidth usage is not abnormal. If you want to limit or reduce it, you can try different settings of the -capacity command-line option. The default is unlimited; you might try -capacity 10 and adjust from there.
The (currently) two Snowflake bridges together transfer 100 GB about every 2 minutes, on average. So while 100 GB in 24 hours for one proxy seems like a lot, it is still just a small part of total Snowflake traffic.
I’m sorry that the proxy is crashing and restarting. It is probably a bug in resource management. You can read more at the below thread and issue: