Server Downtime:
In today's world we need to keep our organizations up and running 24/7 with limited unscheduled enterprise downtime, or face the wrath of IT directors and CEOs. It wasn't always like this, as a few years ago big “mainframes” dotted the infrastructure landscape. These behemoths were very reliable, but required so much TLC that they gave a new meaning to the phrase "high maintenance,” requiring four hours of downtime every third month for scheduled maintenance. So, why aren't we still on mainframes?
Well, most IT professionals have a host of answers to this question including flexibility, cost, and ease of use to name a few. In fact, there are very few System Administrators or IT managers who long for the days of mainframes. Mainframes are nearing extinction, and for good reason. Today's infrastructure environment is much more flexible, with a more distributed architecture and dynamic environments that allow features and hardware to be easily swapped in and out. So, why can't we have the best of both worlds; the reliability of a mainframe with the flexibility, cost, and usability of a distributed environment. Well, we can. The key is to apply the principles of effective server monitoring and to understand the pitfalls that continue to take down systems today. Understanding how to manage CPU and Resource overload, Disk Space, and IT Services in your environment can help cut your unscheduled downtime by 50% or more. To find out how, read on.
Tip #1: IT Services
One of the most frequent causes of unscheduled downtime for critical applications is a critical service stopping or stalling. In fact, a stalled service can be more disruptive due to the difficulty in detecting it. In this example, let's look at a Windows environment. There are three critical applications in Windows that need their services monitored: SQL Server, IIS, and Exchange. While all three are typically used for mission-critical applications, Exchange can be susceptible to one of its services stopping or stalling. The results can be catastrophic, and happens far too often. A four-hour time span of unavailability for a major enterprises email system can be extremely expensive, and cause executive management to start questioning the competency of the entire IT department. Today everyone in an organization, from the CEO to the warehouse staff, rely on email to be effective in their role. People instantly notice when their email is not functioning at 100%.
It is important to note two Windows services properties: Windows services are usually critical and hidden. For example, consider the spooler service. The Service Control Manager's applet interface (the icon is made up of the two intermeshed straight-cut gears) can be used to stop the spooler service. However, if you then try to print you will receive an error message about printers not being installed. Although this is just one small microcosm, and doesn't sound as mission-critical as losing Exchange, imagine the impact of something similar across an infrastructure and business units including HR, Operations, and Marketing.
In a few extreme cases, the production server will need to be rebooted. Fortunately, this is very rare. The bad news is that under 15% of production servers have their services monitored.
Real Life Example #1 - A Large US Credit Union with over 50 Years in business and more than $1 Billion in assets.
Originally focused on purchasing a simple event log monitoring product, this Credit Union quickly saw the huge benefit in monitoring all aspects of their Linux and Windows infrastructure. The key was the ability to manage these systems from one solution, the up.time solution. up.time is used in this environment to monitor various systems including Linux and Windows 2000/2003 servers for Performance Metrics, Exchange, and SQL Server.
In addition, due to management pressure and soon to be in place 'Service Level Agreements,' up.time's Service Level Management features are being used to create report that document server uptime, CPU utilization, and more. In their search for a Service Level Management tool, this client considered purchasing multiple products to meet their service level requirements. However, after trialing up.time, they found that up.time could address all their monitoring, alerting, and reporting needs in a single product.
