uptime of web hosting service.

flowcloud

New member
Uptime in context of Web Hosting Service, is the web server working time without failures. Uptime is important measure, because it indicates web hosting servers and maintenance quality. Uptime is the amount of time that a server has stayed up and running. This is usually listed as a percentage, like "99.9% uptime." Uptime is a great measure of how good a web hosting provider is at keeping their systems up and running. The most basic thing for a website to be found on the internet depends on whether the server where the files are stored is online i.e. up and running at any given time. This is called Server Uptime. You will find that most of the web hosting providers will claim a 99.99% server uptime. Some even claimed 100%. If you have time, search for comments by users of the provider on internet forums or elsewhere online. Or go for the larger more established hosting companies who have built a trusted reputation. Most good hosting companies have multiple server data and power backups. A few also promised 100% and they do that by guaranteeing refund of fees in the eventuality that any downtime might happen.
 
The definition of uptime must be clearly stated.
Anyone can have a server respond to ping requests 100% of the time if the ping interval is every 5 minutes.

For us, we provided a realtime analytics page for not just one server but for a few dozen. Each was pinged every minute, and also showed the following information;
  • Ping Response Time
  • Average CPU load
  • Average Memory Usage
  • Average Disk Usage
  • Average Traffic IN/OUT
  • Last Apache reboot time
  • Last MySQL reboot time
  • Last Server reboot time
We then queried the server with an Apache & MySQL request (search for text response) so that the uptime measured the actual response of Apache & Database as you can easily have Apache running and MySQL be down.

We ran this for about 40 servers and provided the page publicly.

True transparency is hard to find these days - but when presented right, you can win over a customer very easily!
 
I'm always curious about 100% uptime guarantees, because really it's just not practical unless you fudge figures somewhere.

Unintended things do occasionally happen, however good the server admins are and whatever hardware and software you are running. The key thing is to have systems in place to make sure that they are resolved quickly and downtime is minimised.

But even when they don't, sometimes updates or config changes mean restarting servers, it results in a few seconds of downtime but that still means it isn't 100%.

In our SLA we promise 99.98% uptime and most of the time we beat that, a lot of the time we actually are at 100%, but we like to be realistic.

The other important thing is what happens when there actually is downtime, is it just an 'oops, we are sorry' or do you have something in your SLA compensating clients if sites or servers are down and it is your fault. Again we do, and everyone should but a lot don't.
 
restarting servers, it results in a few seconds of downtime but that still means it isn't 100%.
Most companies exclude Planned Outages as part of an SLA. Planned restarts are not usually counted.

Also, most companies exclude Emergency maintenance for hardware or software errors that can be resolved in less than an hour.

The fine print on SLAs are always interesting :)

Godaddy's SLA is 99.9% over a 1 month period. Failure to reach that, you can contact them for a 5% refund of your hosting fee which may be granted at their sole discretion. SLA does not apply to (1) scheduled maintenance or repairs, (2) interruptions caused by users' scripts/code, (3) outages to FTP or email, (4) causes beyond their control, (5) outages related to the reliability of certain programming environments.

It's a pretty blanket "get out of jail free" and if they are liable, they're on the hook for 5% of the hosting fee - only if the user has requested it.

SiteGround's SLA is 99.9% on an annual base. If they are in 99%-99.9%, you get 1 free month of hosting. Also get 1 month free for each 1% under 99%
The following does not count toward the SLA; (1) Scheduled maintenance, (2) emergency maintenance if resolved in less than 1 hour, (3) downtime caused by DNS or IP changes where the user was not notified, (4) dDOS attacks, hacker attacks etc, (5) Downtime caused by user's own configuration, (6) downtime due to exceeding server resources, violation of tos, downtime during upgrade/downgrade of hosting plans, downtime during processing of technical support requests, or events beyond their control. All calculation of downtime is through their own tools and outside 3rd party reports are not accepted as evidence.

Another, pretty blanket get out of jail free as they use their own tools only as the reporting factor.
 
Uptime is the time that shows that your server stayed up and running. This time is usually listed as a percentage; for example, 99.9% uptime.

Uptime is a great measure of how well a web hosting provider keeps its servers active. A high uptime percentage means the server stays up most of the time.

As a result, the site you have hosted there will be available for your visitors. If your website is down, customers can’t keep waiting until it gets active, so uptime is very important.

Sometimes uptime percentage might be misleading. Like 99% uptime sounds great. But it means three days over a year.

When selecting a web hosting service, look at their uptime guarantees; I recommend selecting the one that provides at least 99.5% or higher.
 
The fine print on SLAs are always interesting :)
Very true, and those two examples you cited make that incredibly clear. They are basically saying that almost nothing that results in downtime actually gets counted as downtime.

There do have to be some caveats in there, we have some, everyone does, but those are ridiculous.

I'm curious what others on here include in their SLA about downtime. This is taken from ours at S4:

The amount of compensation will be 30 times S4 Hosting’s charges for services to the client for the duration of the the service interruption. This is up to a maximum of 100% of S4’s charges for services during the 30 calendar day period preceding the interruption per individual interruption. The total amount of combined compensation for multiple service interruptions can not exceed 200% of S4’s charges for services during the 30 calendar day period preceding the latest interruption.

You may check the status of your hosting server uptime from our status page. You should contact our support team if you believe that you have experienced any downtime that falls under this SLA.

The following events are excluded from our calculation of downtime:

  • Scheduled maintenance which clients have been notified of 24 hours in advance, and which lasts for less than 1 hour.
  • Downtime caused by DNS and/or IP address changes which clients have been notified of in advance but have failed to update their unmanaged name servers.
  • Downtime caused by the actions of, or software installed by, the client that is not managed by S4
  • Downtime caused by a client violating any of the General Terms of Service or the Acceptable Use Policy.
  • Downtime during agreed upgrade/downgrade of a clients server resources.
  • Downtime during processing of your technical support request(s).
  • Downtime resulting from law and public authority enforced activities.
Our calculation of network availability is based on our internal and external monitoring systems. Clients do not need to submit third-party reports as evidence to be entitled to compensation under this SLA.
 
Top