Someone can make a concise point out of this error page on my odoo installation?

"nginx 502 :: Bad Gateway"

It happens every now and then irregularely and apparently without further interaction...

Here are the different log-excerpts: (github displays it nicer)

https://gist.github.com/blaggacao/9577fd6b496c9af88735

restarting the odoo service solves the problem until it occurs again.

Update:

It could probably be two problems as error logs are not time exactly related (I don't know much about time dynamics of error logs, but I would supose they should coincide completely or differ only milliseconds, if related)

So the startupscript contains the following lines:

DAEMON=/home/odoo/odoo/openerp-server
NAME=openerp-server
DESC=openerp-server

thus not starting in gevent mode.. (If I'm right...)

On the other hand, ps -efH shows:

https://gist.github.com/blaggacao/e55581a3a2d5503b8273

I configured Workers = 5 (2xcores +1)

You can see, odoo spun up 7+1 where 5 as defined, 1 is for cron jobs AFAIK, 1 I don't know and the last one (the "+1") seems to be something geventish...

My nginx-config btw looks something like this: https://gist.github.com/anonymous/4b3cf30dcc81058ffe1b

Can anyone share his insights/knowledge, please?

I read this one, to no avail: http://www.nginxtips.com/502-bad-gateway-using-nginx/

----------------------------------------------------------------------------------------------------------------------------

UPDATE

I get an intersting new error  in the logfiles, which also has been reported in github.

Something is killing all python instances/threads and leaves odoo down related to an issue with the pickles module.

Reference: https://github.com/odoo/odoo/issues/3878