With Media Temple's Grid service, your website is no longer tied to an individual hardware server. Rather, it is spread across hundreds of server processors. This model gives your site ever-expanding grid-cluster computing capability, which allows you to scale far beyond older shared server systems—and even exceed what some dedicated servers can handle. So your sites are always ready for intense bursts of web traffic.
Grid Performance Units (GPUs) measure the CPU resources being used by each request to your Grid service. Built-in reporting breaks down usage by each hit, helping you track down errors and code inefficiencies. Your Grid service comes with enough GPUs to power 99% of the sites we've ever seen. If you happen to be one of the lucky ones with massive levels of sustained traffic, additional pay-as-you-go GPUs are there when you need them.
A GPU is a measurement based on CPU time required by every single hit/request made to your Grid service. This includes your web server activity, programs, scripts, and all associated parent and child processes across all nodes within our system (100's of processor cores). The Grid system keeps track of all the measurements taken from your server and adds them to a running total, which is calculated about once an hour.
You can view a breakdown of your daily or hourly GPU statistics at any time from inside the Account Center.
Also, if your server is projected to exceed its allocated GPU usage, a courtesy email will be sent to you automatically by our systems.
If you should exceed your GPU allocation, you will be billed at the end of the month for the overage that your server incurred, at the rate of $0.10 per GPU.
However, this is relatively rare. Based on our analysis of over 10,000 random sites, the vast majority (99.97%) would never exceed the GPU allocation, even with heavy traffic. As an example: Even if your service effectively used 2 GPUs per hour during the day and 0.6 GPUs per hour at night, at the end of the month, you would not have gone over the GPU limit.
At the beginning of each new billing period, the Grid system resets your GPU count and begins to calculate a projection of how many GPUs your server will consume. When the hourly usage results multiplied by 24 (hours in the day), multiplied by 30 (days in a month) exceeds the amount of GPUs allocated in your plan, the Grid system generates an email message as a courtesy notice to you. You can also view a GPU report for your server directly inside the Account Center.
In very rare circumstances, it is possible. We have seen a few occasions where customers have reached such severe resource demands that their ending monthly bill would greatly exceed most hosting budgets. In most cases, the situation stemmed not from organic traffic patterns and usage, but rather from malformed software, hacked scripts, or highly inefficient programming techniques.
To protect our customers, it is Media Temple's policy that any Grid service consuming more than 250 GPUs per hour will be suspended until Media Temple can communicate with the customer and determine the source of overly demanding resource utilization. If you would like to pre-approve your service "bursting" past 250 GPU/hr, please contact us so we can set up an exception to this policy.
Not at this time. Grid was designed to handle large bursts and traffic spikes. Restricting its performance is not something that was built into the design of the technology.
No, it’s only a projected overage. The email is letting you know that if things continue as forecast, you may go over your allotted GPUs for this billing period and have to pay for the overage.
$0.10 per GPU
Optimizing scripts and disabling plugins are the simplest steps you can take. Read Understanding the GPU Tool in the (mt) Community for more suggestions.
The moment your service is projected to go over the limit, our system will automatically send you an alert. Any actual overages will appear on your (mt) invoice.
Sorry, GPU alerts are not customizable at this time.
Sorry, we currently do not offer GPU packages.