Maarten Dammers, 01/09/2013 16:37:
Hi Mono,
Op 1-9-2013 16:32, User Mono schreef:
Can tool labs handle this many hits?
Only one way to find out :-) We should probably keep an eye on Ganglia and, if resources really become a problem, request an upgrade of the machine.
You mentioned that the (user?) database on Toollabs/Labs is very slow, is that the main bottleneck? Or in other words, what hosts should one keep an eye on in ganglia? I only see a huge peak of network and a pseudo-swapdeath for a couple days but that could be anything. http://ganglia.wmflabs.org/latest/?r=month&cs=&ce=&m=load_one&s=by+name&c=tools&h=&host_regex=&max_graphs=0&tab=m&vn=&sh=1&z=small&hc=4
Nemo