[tor-relays] Relay overload

[cmeclax]

I noticed when sshing to my relay box that it is responding much slower than 
normal - taking a minute or two to respond to a single keypress. I ran netstat 
-t and saw lots of bytes in the outqueue to various other relays. I telnetted 
to port 9051 and got the following bandwidth log:

650 BW 63449 10568
650 BW 62510 21671
650 BW 63433 22438
650 BW 40314 19035
650 BW 94448 61126
650 BW 10408 11154
650 BW 123157 58027
650 BW 62837 31070
650 BW 12002 3536
650 BW 159449 65423
650 BW 64487 23865
650 BW 65471 35381
650 BW 68892 21466
650 BW 65378 46215
650 BW 38807 11434
650 BW 90042 43329
650 BW 65909 44768
650 BW 77162 18627
650 BW 107393 56922
650 BW 67305 36924
650 BW 25255 11149
650 BW 79619 37266
650 BW 123230 25956
650 BW 67194 27742
650 BW 70871 49787
650 BW 62114 50882
650 BW 53981 60433
650 BW 42699 21631
650 BW 161246 77324
650 BW 66421 39994
650 BW 68306 51876
650 BW 37977 21411
650 BW 157806 71560
650 BW 67964 35298
650 BW 64643 38299
650 BW 41193 26601
650 BW 156710 106887
650 BW 18786 14135
650 BW 158192 48526
650 BW 62862 39633
650 BW 67273 38448
650 BW 57955 39696
650 BW 34695 28979
650 BW 156069 81872
650 BW 53572 28598
650 BW 159163 43722
650 BW 65412 41567
650 BW 67271 48134
650 BW 65372 46107
650 BW 36323 14414
650 BW 0 0
650 BW 158006 92438
650 BW 64455 56813
650 BW 53169 11364
650 BW 109892 82106
650 BW 65191 34845
650 BW 12916 10803
650 BW 76387 21984
650 BW 0 0
650 BW 91819 54140
650 BW 63323 43265
650 BW 66760 26847
650 BW 12030 5817
650 BW 62995 16102
650 BW 154965 70999
650 BW 68408 38650
650 BW 87724 30290
650 BW 44543 45024
650 BW 60544 35343
650 BW 103227 35954
650 BW 155268 53094
650 BW 64928 37615
650 BW 66599 33239
650 BW 111463 36394
650 BW 76657 67405
650 BW 65448 40502
650 BW 65778 49605
650 BW 23300 7620
650 BW 155335 70596
650 BW 65994 31356
650 BW 66109 27590
650 BW 158464 72332
650 BW 60705 36345
650 BW 102858 25936
650 BW 161401 24083
650 BW 86788 62280
650 BW 66162 54081

My throughput limit is 64K, which has not previously been a problem. I've 
sometimes seen steady throughput close to 64K; more often it's sporadic with 
an average around 5K. Any idea what's going on?

I checked Blutmagie; I've lost the fast flag, but have not been selected as 
guard or any other new flag. I thought it might have something to do with the 
bitcoin crash that's going on, but netstat -t on the bitcoin box doesn't show 
any similar attempt to push more data than the line can handle. If I were 
selected as the middleman in one user's file upload, I'd see a steady stream of 
traffic, not this spikiness.

cmeclax