I run about 5 different located servers, through a number of hosting companies. Previously two several weeks, among the servers Sometimes on continues to be lower two times. Both occasions were unpredicted and fairly lengthy (36 hrs and 4 hrs). The server under consideration is really a VPS not really a shared server. Given my knowledge about my other servers/companies (both VPS and shared) this appears as an unacceptable quantity of down time.

  • What is your opinion?
  • What you think an acceptable quantity of down time for the servers (planned and unplanned)?

You receive that which you purchase.

What's your SLA together with your provider? Would you have one? If there's anyone component that describes the main difference in cost, it's this. If you want guaranteed uptime (3 9s for instance) then you will need to pay for this. 5 9s uptime, for instance, can cost you substantially more.

To reply to your question: have you have an explanation regarding the reasons for this down time? 36 hrs is excessively lengthy. 4 hrs not always (if it is rare). Was it a hardware fault? If that's the case, you cannot do much about individuals. I remember when i were built with a provider who does from time to time stuff up their config and mail would cease working until I said excitedly to repair it. In my experience, which was unacceptable.

Server hardware will fail. It is simply dependent on time. Instead of trying to determine which is affordable I'd request you another question: What exactly are all the good ways that the configuration could fail and do you want to improve your setup to take into account these options?

For instance, let us state that your site is located on one VPS. A couple of good examples of failures may be:

  1. The VPS could become corrupt
  2. The hypervisor could fail
  3. Network equipment within the cabinet could die
  4. Energy/warmth problems could appear in the information center
  5. Backbone internet connectivity could drop.

You can decrease your chance of #1 and #2 taking lower your website by implementing a lot balancer an additional VPS. Is decreased risk well worth the additional expense?

This discussion becomes dependent on disaster recovery sooner or later.