Network monitoring updated

May 18, 2007

One of the things that we never want to hear from our clients is that one of our servers or our network is down. Now we know that from time to time there will be network and server issues, but we never want to learn about it from a client. The only reply we ever hope to give is “we know about the issue and will have it resolved soon”.

To that end, we have recently updated our network monitoring systems. We actually have three redundant monitoring systems. Two are internal monitoring systems and one is a “public” monitoring system.

Internal Monitoring

We have two internal monitoring systems based on the open source software Nagios. One system is located in Texas while the other system is located in Georgia. By locating the monitoring “away” from our data center in Chicago we get a “client” perspective of response times etc. All major services are monitored including:

  • Web
  • Mail
  • FTP
  • MySQL
  • Server Load
  • Disk free space
  • All services are monitored every 5 minutes. Doreo technicians are notified if any service fails tow times in a row.

    Public Monitoring

    We feel that it is very important to have a 3rd party monitoring our servers also. This gives clients and potential clients an unbiased view of our uptime. In the past we uses a monitoring service from Alertra. We have recently changed to using HyperSpin. We live several of the features of HyperSpin and were able to negotiate better rates with more frequent monitoring. Our servers are checked every 10 minutes by HyperSpin. HyperSpin utilizes a network of monitors across the globe.

    Finally, you can view the status of our servers and services at http://www.doreo.com/status/ . The status web page shows a “traffic signal” like status for each service. Green means that a service is up and functioning while red means the service is down or not functioning properly. Finally there is a link to the HyperSpin data for each server. This pop up shows the following: