If you are reading this, chances are that you may have received a single or even multiple alerts regarding going over the CPU usage limits. Hopefully this article will explain what is going on and what you should do.
To understand CPU usage and restrictions, you will need to learn more about our VPS/VDS product lines and how they differentiate from each another. The following article will give you a brief summary of that.:
Now that you have read the article, the next step is to understand what is fair share CPU usage policy and how does it impact you.
We believe in transparency and setting the proper expectations upfront to avoid any future confusions. These limits are well defined and outlined within our Terms of Services. You can read more about the limitations here:
So what is "Fair share CPU Usage"?
As outlined under the "Frequently Asked Questions" section, which can be fond here: Nexus Bytes - Premium AMD Ryzen KVM VPS
VPS's are hosted on physical hardware (referred as host node). Each VPS is allocated a certain amount of resources, of which some are dedicated (usually RAM, storage space, IP address etc.), while others such as CPU (vDedicated plans are an exception) and network throughput (known as port speed) are shared.
If the host node is a building, your VPS is an apartment in the building. While you have your own personal space and resources, other resources such as elevators and stairs are shared with others.
Our fair share policy ensures that bad, noisy neighbors will not impact your service. Everyone likes to party once in a while, but no one likes it when that party next door goes on for days.
Our fair share policy gives you the room to exceed your allocated resources for extended period when you need some extra power, but if you constantly exceed you allocation, you will affect other VPS's on the host node.
Our fair share policy allows us to avoid overselling, which in turn benefits you as you do not need to pay extra just because you exceeded your allocation once in a while.
How does these alert work? Why am I getting these alerts?
The way these alert process works is, system would run a check every minute and create an internal Flag for high CPU usage for VPS that are exceeding CPU usage limits. If it flags 29 times (~30 minutes of usage, slightly more than what is allowed by fair share policy), it would send out an email to the user, notifying that their VPS is humming a bit too strong.
Then the system waits again for another 20 flags (Approximately 20 more minutes), before sending second alert and temporary throttling offending VM. So in total, there's ~50 minutes of extreme or high CPU usage (which is still way above our fair share limitation), before any sort of action is taken.
This is a temporary throttle and it is automatically removed once our system detects that the overall usage went down. However, if a VPS is throttled multiple time, within a short period of time, then the VPS CPU usage will be restricted permanently.
This is done to ensure noisy neighbors aren't impacting your regular usage and maximum host node level reliability.
So in total, it waits almost an hour, before any additional action is taken. This is twice the limit that is defined under our fair share usage policy, as outlined above.
- However, there's small logic we use, to ensure people, who are bursting at random don't get penalized. If there's no flag for a smaller period of time, counter goes back to 0 and thus, resetting all the notification and counters.
You can check your VPS resource usage history by Logging into
- VPS Management panel
- Navigating to "Graphs"
- Clicking "System Statistics"
I have been restricted, now what?
So you have gotten restricted? This is not the end of the world!
- If this is your first time getting restricted/throttled: Please wait 90 minutes from the last alert. After that, power your VPS off, wait 10 seconds and power it back for the restriction to be removed.
- If this is your 2nd time getting restricted within the same month, please do the above.
- If this is your 3rd time getting restricted, within the same month, unfortunately this is permanent for next 2 weeks. Please open a support ticket to discuss your next upgrade options.
Quick notes:
If you only occasionally receive these alerts (compiling initial OS, software updates, installation of certain modules etc.), then it is perfectly safe to ignore those. If you have received multiple alerts, we suggest you to optimize your resource usage to be more resource friendly. If you believe that you have done your best, your next step would be to upgrade your plan. You may try by either adding more shared cores or switching over to our vDedicated plan which has no CPU limits thus allowing unrestricted CPU usage 24/7.
If you have any questions or concerns, please do not hesitate to reach out to us via support ticket :)
