I had installed APM v1.0 & SP1 on 7th March
Had a few instances of monitored applications on hosts (very random) timing out or giving incorrect notifcation of being down. Reading a few of the other posts, this seems to be a known issue & I shall leave it for SP2
However, afternoon of the 11th our SQL server for Orion went 99% CPU busy & the APM timed everything out with grey appearing & not being able to contact hosts, the hosts were fine & responging to NPM, although this was sluggish because of the APM hitting the SQL DB so hard, but the older app monitor showed all apps as up etc, all other aspects of Orion in fact were working absolutely fine.
All the hosts are automated trading devices & it was an extremely busy part of the day, but this is normal behaviour at least 2 - 3 times a week, we've previously not seen this type of CPU hogging by an SQL instance, in fact we run 2.6Tb DB with trade search lookups in realtime & the SQL job never consumes more than 30%
The upshot is we've unistalled the APM, also noted that it does not disappear off the website, now the hosts SQL CPU load is more like 8% & all other aspects of Orion seem not to have been affected
Be interesting to see if other folks are running SQL instances for the Orion environment on VM hosts or not, we will most likely move it onto one of the dedicated SQL hosts in the cluster after this behaviour but for the moment & until the odd behaviour of the monitored applications is identified & fixed we are not going to be contemplating this for the moment
thanks
Kieran Omelia
Liquid Capital Markets
Network / Trading Support Analyst
