Burke County Uses Application Monitoring to Troubleshoot Network Slow-downs

WhatsUp Gold application monitoring tools increase network performance and application efficiency for the offices of Burke County.

Western North Carolina’s Burke County covers 515 square miles of some of the country’s most striking countryside, dominated by the massive Table Rock landform. More than 90,000 people depend on the choices IT Director Steve Bennett makes to keep county-critical services like the courts, general services and emergency services operating in peak form across 23 townships.

IT services are delivered primarily to six locations, each linked by fiber in an all Cisco environment headquartered in Morganton. with three or four switches per location. They use Cisco IOS Netflow to each building, part of the county’s effort to maintain very high levels of accessibility for about 40 servers, including Windows applications used throughout the county.

Plagued with chronic application slowdowns

Until recently things were going very well, but as time passed the county’s Microsoft SQL Server and Exchange Server applications were experiencing chronic slowdowns that began to negatively impact county services.

“When Exchange was running sluggishly, it was difficult to figure out the root cause of the slowdown,” Bennett says, “so we would usually start by rebooting the server, which interrupted users. Eventually we’d find that the server wasn’t the problem, that there was something else causing the slowdown.”

Time for better tools

Bennett decided now was the time to acquire a better set of troubleshooting tools than the tools that they were already using as part of the Cisco environment. “We wanted a single pane of glass to monitor all applications,” Bennett says.

Bennett and his team talked to several different vendors before settling on WhatsUp Gold and its integrated Application Performance Monitor. Cisco’s own LAN management product concentrated on network environments, not application availability. And while their server platform monitoring software viewed servers from I/O perspective, it was not set up to monitor the applications residing within them, so they continued their search.

“We looked at another network monitoring product that appeared similar to WhatsUp Gold,” Bennett says, “but they license on a port-by-port basis. Their pricing would be OK if we were just monitoring one or two servers, but we had more than 40 servers and 40 network switches. One switch with 24 ports would require 24 licenses, and this was cost-prohibitive. The same company prices licenses by the operating system as well — each NIOC required a license — further adding to the cost.”

After a full evaluation, Bennett decided to purchase WhatsUp Gold and Application Performance Monitor (APM). He explains his choice:

• “WhatsUp Gold and APM bundle in many functions that others charge extra for. And it prices per device, not per port.
• “Cisco recommended Ipswitch.
• “A single pane of management glass gives us control of the whole network.
• “Support for CISCO Call Manager allows us to use APM all the time inside of the Cisco Call Manager to see problems like a dropped call that can be traced to the telco or network.”

Problems resolved far more quickly with APM

Now, when there is a problem such as the Exchange Server slowdown described earlier in this story, APM gives Bennett and his staff the tools they need to respond much more quickly. “In this example, APM might tell us the slowdown comes from an IIS problem, and resetting it will restore it to normal service levels. If it is a server issue, but not serious enough to interrupt everyone with an immediate reboot, we can simply reschedule it to off-peak hours.”

SQL Server slowdowns can be resolved faster too by using APM as a preventative maintenance tool. Bennett and his team can check trends by viewing easy-to-read graphs, and automatic alerts can highlight issues such as a CPU running hot, disk space utilization exceeding a preset threshold or CPU processes that rise over 60%. “We have a SQL server that has been dropping a database a couple times a week, so we’re putting APM on that server to see what issues affect the SQL database. We will look at things like utilization on server disk and memory. Perhaps there is a particular service giving it a problem. I can look at servers that run multiple apps and use APM to determine which app is giving me problems.”