Search User Profiles

Search All Sites

Login

Remember Me
New User? Create an account

Comcast Featured

Comcast


Bookmark and Share

Organization Type
Corporation
Industry
Media (Publishing / Broadcasting)
Country
United States of America
Hits
73388
Infrastructure Notes
2 servers: each has a dual Intel Quad Core CPUs, 16 GB RAM, ~ 40 GB SCSI disk + Gigabit NIC. 1 server is dedicated to polling 1 and the other server is dedicated to reporting * PNP with rrdcached for short term trending (up to 14 days) * In-house data warehouse for long term trending (based on Vertica) * status2json.cgi for external JSON feed * Custom notification system built with a perl dispatcher and a ruby front-end (will release GPL as time permits) * SNMPTT for trap filtering and a customized Nagtrap-based GUI for trap viewing and searching
Monitoring OS
Linux (RedHat)
Monitoring Setup
  • Standalone
Nagios Servers
2
Monitored Hosts
1600
Monitored Services
8900
We pass all data to a data warehouse for long term trending, so completing each polling cycle in 5 minutes is critical for us; Nagios lets us do a relatively high volume of checks per minute on our poller (~ 2200/minute with our current hardware). We really enjoy working with Nagios and thankfully are able to contribute modifications and patches back to the Nagios community.

80% of our checks are run using perl plugins under ePN (90% SNMP, 10% other), 20% use NRPE.

Special thanks to my managers Mike Fisher and Jason Livingood for allowing us to contribute code back to the Nagios community and share our experiences with Nagios with others.

We are just about to expand into a second set of servers and will most likely use DNX for distributed monitoring; if we go with traditional distributed Nagios we will use the Bronx NEB module instead of NSCA.