Metrics for NOC and network guy
Hi Folks,
Customer here only eval DCRUM Enterprise Decode, they have no interest on HTTP, HTTPS
May I ask any suggestion on what kinds of performance / availability metric to convince customer (system guy & network guy) ? I tried to move them to application approach but was refused. They want troubleshooting on telling performance / availability of network & system.
I introduced metrics such as usage, RTT, response time, TCP errors....etc, anything else that can attract system guy & network guy?
many thanks.
Sylvian
Hi don't forget that you have basic FDI from a network level view as you know if the RTT or ERROR came from the inside or from the outside.
You will also gain insight into whether they have Jumboframes and legacy protocols in the network such as Wins and what not. At te same time you get clarity on protocol distribution and potential rogue processes such as backups being done during production and AnitVirus scan/updates that use more bandwidth than what is sensible - all of these eating up what's there to support business applications.
I had one installation where the SNMP poller of a system monitoring tool ate +30% of the WAN - leaving Citrix in problems due to the burstiness of the SNMP poll and the lack of QoS for the various channels in the network.
