<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> Yeh, I read about that here on the list, so I started the tests without the perl. Your load is really strange, it seems that something is running at a predefined intervals, something in the cron? I also had a graph to measure memory utilization, rss, size and vsize, of the nagios process and there is no change in this, although latency keeps going up. At a previous test, the latency reached 800 seconds and was still going up, and the load started to go down, we might see this soon in the running test systems. As I had not explicitly added to the configure the disable for the embedded perl, so I reset the test again 3 days ago, but I am getting the same problems. A nagios restart will not bring the latency down nor it will stop it from growing. Another thing that may or may not matter, I am using the same configuration for objects as in nagios 2.5 and I have enableed the large installation tweaks in the two nagios3 boxes. If there is anything I can do to give the list more information, just let me know. []s. Gerd Mueller wrote: <blockquote cite="mid:1202916188.27735.3.camel@netl-gm-01.int.netways.de" type="cite"> <pre wrap="">ups, I must have missed the "not" :-( sorry. A few days ago a perl script caused huge latency on my installation. That is why embedded perl is one of my usual suspects ... My nagios' latency is pretty good ~ 0.4 seconds but the load on the the system is pretty strange (have a look at the attached graph). Anybody any idea? Cheers, Gerd </pre> <hr size="4" width="90%"> <center> </center> <pre wrap=""> <hr size="4" width="90%"></pre> </blockquote> </body> </html>