Why monitor your servers? As with all resources, computer resources are limited. That means limited work can done by the CPU, limited information can be held in system memory for use, and limited data can be stored on disk.
When a system has all of the resources it needs, it generally runs fine. But what are the costs when a resource is depleted?
Strained CPU and Memory Resources
Running out of CPU processing power (using it at 100% capacity), or running low on memory, means work that needs getting done is delayed.
Costs:
- Sluggish or non-responsive websites, which leads directly to customers leaving.
- Failed or timed out database transactions, which means failed orders.
- Company knowledge workers end up waiting for reports or other data, meaning idle human resources.
Running out of Disk Space
Running out of disk space might be the worst because restarting the computer might not help anything.
Costs:
- Databases cannot record orders, inventory changes, payments, etc. Orders are lost, business data is not reliable, customer information is not recorded.
- Audit logs cannot be written which translates into failed internal security controls, lost compliance, and somebody being in hot water during the next audit.
- Emails cannot be received and stored in the mail system, leading to failed communications, lost orders, bad customer support, and a grumpy CEO (we’ve heard stories!)
Going Deep
The above examples are the simple ones. There is so much more that a good monitoring solution can do to help. For example, knowing you have enough CPU, RAM and Disk is a great start.
- Would you like to know if your website is responding sluggishly?
- What if your website SSL certificates are about to expire?
- Should you know if your mail server has crashed?
- How about the database transaction log is almost full?
A good monitoring solution will watch all of that and much more, and summarize it all into a simple to view dashboard showing you at a glance if everything is OK, or whether some component in your business needs attention.
Each of the above problems are usually predictable, meaning statistics about usage can be measured, stored, and charted to help you see if your system is getting near its resource limits. This could be done manually by checking on those statistics throughout the day, and for just one or two servers that might suffice (though it will surely not catch temporary problems, especially those that happen after work hours). A better solution is a software product that can monitor once every few minutes, or even more often than that.
Over-provisioning
One overlooked benefit of monitoring is you can right-size your computing investment. Monitoring can help you with understanding what resources you need, including at peak times, so you don’t pay for more resources than you need. Having 100 serves all running at 5% usage rates is wasting energy (to run and to cool), indicates too much was spent on hardware, and could just generally be utilized more efficiently. All of that translates directly into saving money, and usually saving much more than a typical monitoring solution would cost.
Savings:
- Only buy/pay for what you need = more money in the budget!
See for yourself!
If you would like to avoid the problems and have all of the benefits of a powerful yet easy to use monitoring system, take a look at PA Server Monitor!
