New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
It would seem after trying their best to ignore this, they think that more lies are going to help them save face.
I use dozens of providers on LET and very rarely call one out, no matter how many billing issues, provisioning errors, how much instability, downtime, severe performance issues, and how slow/incompetent support may be.
The only thing I expect from low end providers is that they are honest, respond to tickets (even if it takes many weeks), and make an effort to resolve issues (even if that takes months). If they do that, I'm willing to gamble that a provider may eventually provide stable services that are cost effective long term, and that gamble often pays off even with some of the lowest tier providers.
This provider has utterly failed on every one of those points, despite being given months to respond to my ticket, and instead of being honest, they simply double down on the lies.
I will have to set the record straight.
I'm not sure how more than 100% CPU usage could even be possible, but regardless, this is blatantly false.
I clearly stated in my ticket that CPU usage of the server was under 60%, as could be seen on your own panels CPU graph.
I host a wide range of services and having CPU usage above the terms of any given provider would risk the server being throttled (or a suspension) and cause degradation of services for my users, which is why I monitor resources and ensure that resource usage is within the terms dictated by the provider.
Too bad I didn't call you out months ago when I could have shown screenshots of your own panels CPU graphs to prove that CPU usage was under 60%, instead of waiting months to see if you would ever respond to my ticket.
I actually opened a ticket before you limited the server, due to steal caused by an overloaded host with my server having under 60% usage, and you responded by limiting the server to 30% CPU.
Your only response to my ticket was initially asking what the server was being used for, and I in fact went into detail about the reason I purchased the server you advertised as having dedicated CPU resources, the nature of the compute workload and services being provided for my client.
You never responded after that, and instead manually limited the server to 30% max CPU usage, making the server completely unusable for production.
From the LET offer thread:
From this new offer thread:
Combined with your terms not stating any limits on CPU usage, you clearly offered dedicated CPU resources where up to 100% usage is allowed.
Go through the recent posts and reason should be clear now.
The offer states resources are dedicated
From support ticket
The panel does not match on memory usage and storage used. Also, cpu usage > 100%
Look at the screenshots
OP should just change wording in description and make everybody happy.
@Cybr @FameSystems
First to show any proof wins.
Technically this kind of emulates dedicated experience quite nicely, if they really don't oversell.
Right now
with top shows zero cpu steal and 100% cpu usage.
I agree, but the offer should clearly mention that cpu vCores are shared and managed dynamically through a load balancer part
Also, cpu usage at 100% is not permitted (confirmed by support ticket). The dedicated term implies no restriction on the cpu usage.
VDS == no restriction on cpu usage
VPS == cpu usage is limited
You mean constant or prolonged 100% is not permitted?
Idk, at this point in time this deal is very good, miners and LLM kings thinking they can bash 12 9754 cores 100% night and day for 20 bucks are a bit naive. Let's see how this holds in two months, when nodes are fully populated.
Proof of what? That the server didn't use over 100% CPU usage?
This all happened months ago and their panel doesn't have a long term history to be able go back and take screenshots, so it's way too late.
The only time the server would have had even 80% usage briefly was when I ran YABS, which was either before or after I left it idling for weeks right after getting it.
Refer to this post
https://lowendtalk.com/discussion/comment/4039552/#Comment_4039552
But support states otherwise.
maybe over 50% cpu long time usage is not permitted, but not 100%
The hard disk usage does not seem very serious to me
I just used it at 100%, so 'CPU (usage) limit is 100%' indeed. They didn't lie.
if so, that is not dedicated resources as famesystems has said in the server description.
so stop play words game
The panel numbers do not match
Idk, maybe start with what "compute workload and services being provided for my client" were you running to achieve a "a CPU usage of over 100% (regularly around 110%) over a period of multiple days"?
That's a "maybe" territory. But I'd say as long as they clearly tolerate people watching software-decoded YouTube over RDP on Windows, it's highly unlikely.
Either they have trouble reading metrics from Windows or Windows just has some CPU overhead that is invisible from the "inside". I've quickly checked it on Debian, 1 core load, all cores load, seems okay-ish, except 'RAM used', that one is way off.
How would you explain this number 107%
How can cpu usage exceed 100%?
unless the software used for collecting the metrics is defective
I have seen that wrong reporting issue with VirtFusion too...
Inside VM - CPU Usage is low
On Dashboard - CPU Usage is High
Theoretically it could include cpu usage needed for virtual network and virtual disks.
They are probably using proxmox.
Proxmox organizes guests always in a cgroup to make sure that a guest can't consume more then what's assigned to him.
If you attach disks with an extra io thread to your vm you technically have access to eg. 4 threads and one further thread which is only used for your guests io.
These io threads are placed into that cgroup as well as backup jobs of vms eg.
You could technically have 300% CPU usage on such a Proxmox VM with only one thread.
You just get the usage of that cgroup reported.
I thought it would be okay to enlighten the community a little bit on that topic.
Just remove my comment @.mods if it isn't allowed to post under other providers threads.
There is something wrong with how panel computes the resources used
Left side -- Windows running in idle mode (except RDP) -- 5% cpu usage
Right Side -- @FameSystems panel reporting 37% cpu usage
We know that our graphs are way off. I personally just know, that they are off when using Windows, since we use Proxmox. And if you use a Windows ISO, it's even worse. We will try to fix it, so every customer can see their real usage in the webinterface. But I don't know how greatly we can improve it. Also, the RAM usage is due to the RAM of the servers being mostly cached. In Proxmox (and our webinterface) it seems like it's using the whole RAM, but it should be just cached.
Regarding the 100% limit. We don't have a limit like 50% over a long period. It's more than 100% for a long period of time. Personally, I don't know why a 100% limit is a problem, it's everything you pay for, we just don't want anyone to use his resources in an abusive way. If you mine for example, and put immense loads on the system, it can reach more than 100%. That's where you will be limited. Especially if you don't have a good reason. And I don't mean over 100% in our Interface, since we know, that these graphs are off most of the time. And you definitely won't get limited if you use over 50%, I don't know why you should, we never did that and never will.
You get your full performance you pay for, without being limited, especially if you have a good reason.
I know it's confusing that we say dedicated, and I don't know why my colleagues always say dedicated, it's not really dedicated, since it's shared. Even with our load balancing, it's not dedicated. But to be clear, you will get your 100% of performance since we ensure a max usage of 50% per host system. In the future, we won't call it dedicated, since it simply is not.
We strive to be as transparent as possible and calling our server resources dedicated is definitely not transparent, and I apologize for that. I hope I was able to clear up some misunderstandings. If you have other questions, feel free to ask.
This is greatly appreciated, @FameSystems
I must also add, the FameSystems support has been very professional and quick to reply all my questions.
Thank you.
Also, any plans to support Windows + Linux without swap the offer? ISO mount will be supported?
yes pls stop the trend of hosts calling it dedicated when it's not a dedicated server.
ISO mount currently is supported if you open a ticket and ask us, we will enable it.
Swapping is a good idea, we currently don't have it implemented, since our Linux and Windows packages never were the same. We only added the same packages as Windows for the sale. I will discuss this with my colleagues, but we will probably add it.