Keeping everything running smoothly is 99% of a server administrator’s job, whether that’s preventing outages or making sure productivity doesn’t slow down. You should be constantly scanning for any glitches, and ready to troubleshoot and solve when they’re found.

You also need to make sure preventative measures are implemented, and act quickly when any breaches or violations occur.

The core tasks of any server administrator are:

• Uptime monitoring – keeping track of what’s up and what isn’t
• Performance control – seeing how well things are performing
• Resource management – making sure everything has everything it needs
• Security policy maintenance – stopping any breaches before they happen

Deciding which metrics need to be monitored depends on your server’s purpose. A server that just holds your public-facing website is very different from a server that holds your files which is very different from a server that holds your CRM software, emails, intranet, and public-facing website.

In most situations, the metrics you’ll need to monitor are:

• Server availability and response
• Performance spikes in the server CPU
• RAM space
• Storage space
• System log files
• DNS performance issues
• State of the components and equipment
• Environment (including temperature) of the physical location

Luckily, you don’t have to take all these metrics by hand. As servers have become more and more connected, and applications more complex, automated monitoring has expanded.

With an automated monitoring system in place, you can review reports, graphs, and alerts that let you know what’s happening in real time. And with those details, you can optimise performance, map out potential problems down the line, and save time and money.

Because monitoring is so important, you’ll need to make sure you think carefully about your choice in an automated monitoring system.

Cost is important, but other factors have to be considered:

• Are you planning to upgrade your systems any time soon?
• Will it be able to cope with scaling up or down?
• How flexible is it?
• How simple is it to understand?
• How much will it free up your time?
• How does it list errors?
• If you’re running your website or email through your server, how well does it monitor traffic?

And whatever you decide, always go for one that lets you try before you buy. Along with seeing how well it works with your procedures, it also shows how well it responds and adapts to systems.

 

What are your tips for server monitoring?