There are more uses to these proprietary server analytics than simply ensuring that you’re receiving all of your critical system resources. Let’s take a moment to look at some of the most common ones.
1. Checking your site performance
The stated purpose of the Site Performance Guarantee resource graphs is to show that you are receiving your promised resources. If you see your site performance degrading, it’s very simple to check your resource graphs and make sure you are receiving 100% of your critical server resources.
Resource degradation, though, is so unlikely that not only could we not find an example of it to show you, we could not find an accurate way of artificially producing the problem for illustrative purposes.
The important point to remember when checking for hardware degradation is to make sure you look at the “Allocated” number in individual resource graphs. If that number falls below your guarantee for even five minutes, you qualify for a refund of that entire month of service.
The important point to note here is that for each of our guaranteed resources — CPU, RAM, disk space and IOPS — if your site slows down and you do not know why, you can check these resources at any time to confirm that it is not a problem with the hardware. If it is — you get that month of hosting absolutely free. If you want the fine print, you can read our entire SLA here.
2. Monitor your total server uptime
With ServInt’s Cloud VPS, full server downtime is very unlikely in all but the most dire circumstances. Still, let’s look at what a full server outage would look like in your resource graphs.
We simulated an outage on August 25th by rebooting a test server. Notice that the lines on the graph simply disappear. If you see this, click into a specific resource graph (such as CPU):
You can see in this graph that not only has the resource utilization gone to zero, but the allocated amounts have as well. This indicates that the server itself was offline.
In the extremely unlikely event that you are temporarily unable to access the customer portal, please know that these graphing tools will continue to collect data. This will be valuable after the incident should you decide to seek a refund for downtime.
3. Monitoring resource usage.
Confirming that your server hardware is running at 100% efficiency and your server is up are both important, but frankly, we don’t expect many customers to find evidence of resource degradation or downtime.
What is far more common is for customers who are experiencing site slowness to log into their resource graphs and find that their server is online and they are receiving all of their promised resources, but that they are using all the resources in one or more categories. Let’s look at some examples of this.
These examples show customers who are maxing out IOPS, CPU and RAM respectively. Simply put, if you are using all of the resources your server has been allocated and you then wish to use more, your site will slow down. The question of how to fix this is one of time or money: should you optimize your code or purchase more server resources.
Sometimes customers only have spikes in their resources. These spikes may have been anticipated or may be a surprise. The questions to consider in this scenario are: Did you have enough computing power in reserve to handle this extra traffic, and was the traffic spike preventable or simply a result of increased visits to your site?
Finally, there is a special circumstance when monitoring RAM usage on your Cloud VPS. There are two separate RAM metrics that add up to Total RAM: Active and Cache RAM. Active RAM is just what you would expect — this is the RAM the processes on your server are actively using.
Cache RAM is an optimization wherein unused RAM is temporarily used to store (or cache) data that has been recently accessed on the server. When that RAM is needed for Active RAM, the cache is simply dumped (or a portion is dumped) to free up the RAM. Some servers have a large cache because of many active users. In this case, the Active and Cache RAM total nearly all of the allocated RAM. But other servers do not require a large cache or the cache has been recently dumped due to the need for more active RAM.
If your active RAM usage increases to near the limits of your allocated RAM, you will have little-to-no space left for Cache RAM. Depending on your configuration and requirements, this could signal a need to optimize your code or upgrade your hardware.
4. Planning for future optimizations or upgrades
One of the best ways to use your Cloud VPS resource graphs is to plan for future growth of your sites. Monitoring resource usage over time and deciding on predetermined limits past which you will upgrade can save you the headache of running out of resources in the first place.
Study your resource graphs. Determine how much of your system resources you’re using on a regular basis and how much you need to hold in reserve for growth or spikes in traffic. This will help you know when it is time to scale before you run out of resources. With ServInt’s Cloud VPS, powered by SolidFire, knowing when to upgrade your resources has never been simpler.