Do you know that for a fact, or are you just guessing. How do you know? Because your host tells you? How do they know? What's their monitoring level? How often? A "ping" test every 5 minutes is hardly monitoring. A server could easily re-boot within a 5-minute period, and you might never know it was down. A server can easily respond to a "ping" but not be serving web sites. PHP, MySQL, .NET might all be down individually while other services are working. You have to be monitoring other processes and what do you then base your "uptime" figure on? What if the server's working and serving sites, but CPU usage is sitting at an uncomfortable 95%. Do you include this excess CPU usage as "downtime"? What if your DNS servers are responding to pings, CPU, RAM, and Disk I/O is fine, and it's even resolving most DNS queries but not all. Is that downtime? HTML sites work fine, PHP sites work fine, but ASP.NET is not responding? How do you calculate the "downtime"? What if the FTP server is down so you can't upload new files, but everything else is working, would you consider that "downtime"? Email is fine, but none of the backup email servers are working, do you count the downtime that doesn't have any direct impact, but is part of your overall service? "Uptime" is not a simple figure because the answer depends on what you're measuring and whether that measurement means anything.Is the only true downtime when the server has stopped working altogether, or is it more subtle than that? What about the "fudge factor"? What's their alert thresholds. What happens if a test fails? When a monitored service fails, with some systems it isn't considered to be truly down unless there are 2 or 3 consecutive failures - giving a potential downtime of 15 minutes (or any other period) before anyone notices or does anything about it. They can't respond to every single "fail" because there can be all sorts of reasons for a fail without the site or service actually being down. Support teams will always have some sort of confirmation in their monitoring before they respond. What are you basing their uptime on? We closely monitor all our servers, but I couldn't tell you our average uptime without doing some calculations - or at all unless you tell me what it is you consider to be "downtime". If they only have 1 or 2 servers that might be fine, but what happens in our situation where there's multiple web servers, separate DB servers (8No), 4 dns servers, 4 email servers, 2 monitoring servers, 2 ftp servers, 2 control panel servers, multiple VPS host servers, 3 backup servers, a statistics processing server, a support system server, a customer portal server, etc etc...you get the picture...all on separate hardware. What, exactly is being measured? What if they show a 100% uptime for a web server but a crucial DB server is down and some sites aren't working...but the web server is up and serving sites. I can give you accurate data for individual servers, but calculating the overall uptime depends upon what you're measuring and how you want it presented, and providing "web site uptime" is even more difficult. Is the "ping" response enough? What about PHP? or site response times (we monitor how fast they respond and send alerts if the monitored site responds slowly). How is the uptime being measured? How is that figure arrived at. And what about scheduled maintenance? Is that counted in the overall uptime? Does the company carry out any maintenance? The point is, it's not easy to tell WHAT the uptime really is, coz it depends WHAT you're measuring, HOW OFTEN it's being measured, the ALERT THRESHOLDS, and how services are distributed in a set up and their relative impact. Unless your sitting watching your site 24/7 then you're not going to know the sites true uptime, and unless you know what your measuring at what interval the results could be meaningless. We monitor our servers not so that we can boast about our uptimes but so that we can respond quickly to any problems detected. With so many servers you can't do it manually so you have to have automated systems in place, and you need to be monitoring all sorts of pararmeters on each server to get a good overall picture of what is happpening. A "ping" or "URL" test is never enough, it has to be more detailed and thorough than that.
no need to guess mate. register with website monitoring services such as pingdom.com. they monitor your site's uptime + many other things. they give you regular reports on your site uptime and also informe you everything your site goess down and back up again even for a short 2-5 mins. for the past few month i only had 5min downtime on 1/2/2012 which was due to a backup process. I wont pay attention when a host says 99.9% uptime and practically monitor my own site.
No, I think we need explain a question. Highest uptime can be 100% if you 're on cloud server or if you see a good server which offer nightly maintence mode without restarting anything. Also, I think many hosting providers always to keep their guarantee 99,99% uptime. It means every month they can take maintence mode (with downtime) to 35 min.
I think you might need to recalculate that. 99.99% uptime is slightly under 4.5 minutes of downtime in a 30 day month.
providing true 99.9% uptime is the biggest challenge for every single hosting company. even if their prices are dam cheap or they provide 1GB dedicated server at dirt ccheap price if they cant keep up 99.9% uptime they will always be known as tier 3 hosting. 30 during a month is kinda too much but 5 min during 3 month is more than acceptable.