Hi, even though Icinga should not be your bottleneck (from the stats, and see remarks about RRDtool in previous mails), you could try to set the following in icinga.cfg
auto_reschedule_checks=1 cached_host_check_horizon=150 cached_service_check_horizon=30 This is what I have configured on my similar server with 335 host 2300 service checks without load problems and having also RRDtool updates (<1000 files) on the same system. How do you make the RRDtool updates? Are they triggered at the same time (maybe by Cacti poller)? However, you should at least check the load of your server with Icinga (check_load) in order to have a clearer picture. Cheers Ralf ----------------ursprüngliche Nachricht----------------- Von: "Tomas Macek" [email protected] An: [email protected] Datum: Thu, 6 Jan 2011 08:26:31 +0100 (CET) ------------------------------------------------- > Hi, I have Intel server with 2 xeon 3.00 GHz physical processors (together > 4 cores) and 16 GB RAM on RAID 1 array. I monitor with this about 1600 > services and 1000 hosts. The service monitoring consists mostly (maybe > 95%) on check_ping and check_snmp services. > > I have an experience from the past, that when the nagios/icinga load is > too heavy, some checks are somehow skipped and I can see it for example in > "Host > problems" in the column "last check" - the last check is for example some > hours old, when it should be checked every 5 minutes. Forcing the check > solves always the problem. > For example this morning some host was down but never recovered, altbough > the check ping service on this was OK. Forcing the host check resolved > this. > > Do you think that this hardware is enaugh for such a load? Don't you think > that I'm doing something wrong? Thank you for experiances > > Regards, Tomas > > > -------------------------------------------------------------------- > ---------- > Learn how Oracle Real Application Clusters (RAC) One Node allows customers > to consolidate database storage, standardize their database environment, > and, > should the need arise, upgrade to a full multi-node Oracle RAC database > without downtime or disruption > http://p.sf.net/sfu/oracle-sfdevnl > _______________________________________________ > icinga-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/icinga-users > ------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl _______________________________________________ icinga-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/icinga-users
