Sorry to bring back painful memories: how often did your server go down last year? And how often should we expect them to go down?
Sure, nobody wants their server to go down at all! But, back in the real world, servers do go down. Few organisations have the resources or the luck to maintain 100% server availability. (Think about your car.)
So we analysed downtime on 100 servers, and we split them into two types of problems to show you what server availability to reasonably expect.
Minor Problems (such as your server rebooting at THE most inconvenient time)
Question: How often should a minor problem happen?
Answer: Up to 3 times a year.
Why? Because the amount you’d need to spend to guarantee 100% server availability rarely justifies the hour or so you need to be able to use email at midnight on new years eve.
But there are some provisos. We’ve listed the main ones below.
Major Problems (such as the server rebooting AND then not coming back up at all)
Question: How often should a major problem happen?
Why? Because who can afford to be out of business for a day, or lose client data or order information?
8 Provisos About Server Downtime:
1. Just like your car, regular maintenance keeps servers healthy. Healthy servers get monthly checks, just like responsible car owners do for oil, water, and tyres.
2. Monitor utilisation versus resources. Once your server is 90% utilised, it’s unstable. Imagine driving your car in the red zone on your rev counter all the time!
3. Diarise your hardware warranties’ expiry dates so you’re never without cover.
4. Backup your critical data on-site and off-site. It’s easy to automate backups. Do it once, and make someone (either in your organisation or your IT company) responsible for regularly checking it’s switched on. Then relax.
5. Schedule a check before public and bank holidays. Everyone knows that’s when the mysterious “holiday phantom” uncannily wreaks havoc.
6. Keep tabs on whether your support company told you, or you told your support company that a server went down or is about to go down.
7. Check your IT support contracts to see if your IT support company actually benefits from your servers going down. WTF? How can an IT company benefit from your servers going down? Simple: hourly billing. If you want to incentivise your IT support company to keep your server up, switch to a monthly retainer with an SLA.
8. Dig out your Disaster Recovery (DR) plan. What? You can’t find it? When did your IT support company give you the DR plan? If your IT support company hasn’t asked you about disaster recovery or business continuity, consider that a warning light. Pull over and look under the bonnet of your IT support systems.
Share your own ideas in the comments below. And send a link to this article to anyone whose life depends on their server not going down.