Monitoring a computer network requires thinking and planning before implemanting a monitoring system.
Think about what to monitor
You could start monitoring every aspect of your network, but does that make sense? Do you really need to know if a printer is up every minute? Or is a check one an hour enough. Is checking the CPU load on you main server once a day enough or do you want to keep that und a close watch?
Or do you want to know the status of every workstation on your network? I don't!
You need to make a chart with the most critical elements of your network before you can define the probes that you need to check.
Placing a Time Trigger
Along with the definition of the item you want to track you should write down how frequent this check need to be done.
As I mentioned earlier, you might want to check you main server CPU load. Now that is something that can be triggered every minute or every five minutes so you can get a better view and insight of the load that this server has to endure during normal working hours. Getting such insight will prevent you (as an example) from going on a wild chase if someone reports a server response slowdown at 9,00 o'clock in the morning. You already now from an earlier investigation that at that time a SQL Import is done by the Controlling department to update their credit control program. That previous investigation was done after your saw that peak reoccurring every day at the same time.
Checking your Reports
In the example give before you noticed the CPU load on the same time every day and you could find what was causing this event.
Finding those anomalies is of course only possible if you read the reports and check the screens on a regular base. Make sure you do this at least once a week to see changes and issues before your users do so you can investigate and solve any problems before your users come knocking o your door.
Leave a Reply