Here is a console output:
si14@si14-work:~/repos/monitoring/root$ python2.7 server.py 

127.0.0.1 - - [2011-01-28 12:29:30] "GET /update HTTP/1.1" 200 320 "-" "Python-urllib/2.7"
{"seenby":[1],"received":1296207058.993983,"observer":1,"type":"ping","source":102,"time":1296207058.990101,"data":[[1296206970.543701,0.010154962539672852],[1296206980.383922,0.010203123092651367],[1296206990.222841,0.01015615463256836],[1296207000.050695,0.010264873504638672],[1296207009.876834,0.011881113052368164],[1296207019.698611,0.010120153427124023],[1296207029.519147,0.010251045227050781],[1296207039.342266,0.010113000869750977],[1296207049.167352,0.010238885879516602],[1296207058.990089,0.010174989700317383]],"class":"statistics"}
127.0.0.1 - - [2011-01-28 12:30:59] "POST / HTTP/1.1" 200 2 "-" "Python-urllib/2.7"
Segmentation fault
si14@si14-work:~/repos/monitoring/root$ 
I'm not sure that this is a gevent issue, so I'm posting it here. server.py sources are attached.

I should also say that this bug appears at first time, but it doesn't make it less scary. The packet that crashed python was the same as the one shown and I've already used this tiny "server" for a day or two without modifications, so it seems to me that this is not my mistake.

Ok, I've played with this some more and got some more segmenation faults. It looks like a gevent problem, but I think that python shouldn't completely fall so easy. Here is more traces:

si14@si14-work:~/repos/monitoring/root$ python2.7 server.py 
{"seenby":[1],"received":1296208139.606481,"observer":1,"type":"ping","source":102,"time":1296208139.603046,"data":[[1296208051.083743,0.010155200958251953],[1296208060.923999,0.011048078536987305],[1296208070.76751,0.010570049285888672],[1296208080.613247,0.011930227279663086],[1296208090.454012,0.010123968124389648],[1296208100.283144,0.010128021240234375],[1296208110.114118,0.010215997695922852],[1296208119.943081,0.010147809982299805],[1296208129.774567,0.010593891143798828],[1296208139.603033,0.010541915893554688]],"class":"statistics"}
127.0.0.1 - - [2011-01-28 12:48:59] "POST / HTTP/1.1" 200 2 "-" "Python-urllib/2.7"
Traceback (most recent call last):
  File "evhttp.pxi", line 473, in gevent.core._http_cb_handler (gevent/core.c:13165)
Segmentation fault

127.0.0.1 - - [2011-01-28 12:47:30] "GET /update HTTP/1.1" 200 320 "-" "Python-urllib/2.7"
^CTraceback (most recent call last):
  File "server.py", line 34, in <module>
    gevent.wsgi.WSGIServer(("localhost", 8020), printer).serve_forever()
  File "build/bdist.linux-i686/egg/gevent/baseserver.py", line 183, in serve_forever
TypeError: raise: arg 3 must be a traceback or None
Segmentation fault
si14@si14-work:~/repos/monitoring/root$

I don't reproduce the issue. Could you run python under gdb and give us the C backtrace?

I've changed gevent.wsgi server to gevent.pywsgi and it works as expected. Now I switched it back to gevent.wsgi and it doesn't crash too! That's strange, but I understand that you can't fix it without normal backtrace. I'm sorry for the inconvenience.

Closing as invalid since gevent seems to be the culprit of the segfault.