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
@Hetzner_OL
Why there should be always bad? No, you are simply on the node which has less yabs tests running constantly. Personally I have yabs on cron.
I have read some reddit post saying the same thing. I don't think it is ever possible to achieve 0% at all times unless they gave you full dedicated cores, even my 7900 VPS will fluctuate to 0.1% at times.
The only way you will know if they disable it or not is to run GB6 bench on the moment you feel performance is meh.
I doubt it's disabled. Steal time really should be 0 unless the node is considerably oversold.
It always has been...
Yes, they have. There was a forum post about it, Intel CPUs can only burst up to 20% of available CPU power constantly. For example with the CX22 you only have 0,4 vCPU threads when running the CPU at 100%. For the shared AMD line it's 33%. Only CCX has dedicated CPU. You can see if you get throttled in the Hetzner UI Graphs, it will get stuck at a certain treshold.
Hello, would you like to send me your customer number or a ticket number in a private message? That way i can take a look at your case.
RIP your accounts getting closed now @EthanZou
Guess @EthanZou better have a backup plan. 🤷 Just kidding, @allthemtings but while entertaining us, maybe sprinkle in some actual facts next time? You know, for balance.
What happened to --Katie
Did you ban her account too?
Can you please link the forum post. Not that long ago someone on Reddit commented that they contacted Hetzner, which responded saying that they allow an 80% average utilization. So it's pretty bad if they've restricted CPU usage that much.
Edit: Also @Hetzner_OL could we get some confirmation? There is nothing in your Docs about average CPU utilization allowed in your shared cloud servers. Most providers do explicitly say the CPU usage limits on shared cloud instances, which makes it easier to know if you should upgrade to a dedicated vCPU plan. Thank you!
Oh, Katie? Don’t worry, she’s very much around—still part of the team and keeping an eye on LowEndTalk, probably reading this thread as we speak, sipping her coffee, and deciding how much sarcasm to add to her next response.
-Lea
Who knows, she might even reply if the mood strikes!
I might be wrong, but long, loooong ago (2014 or so) there were guides on how to check how many neighbors are on the same node with you (i.e. "overselling metrics"), whether the CPU steal is real or disabled/fixed, etc.
Those guides were for OpenVZ though. Is something like that possible these days?
We want Katie back to address to the direction Hetzner is going forward. Too much negativity from Hetzner in the recent times
Probably you're using a plan with shared vCPU, the compute resources are distributed among all instances on the same physical server. These instances provide a baseline CPU performance and have the option to temporarily burst CPU usage above the baseline level. And this allows us to provide these shared plans to you at a much lower price.
To ensure a fair allocation of resources we use an active management system to prevent resource shortage. This system temporarily limits instances which have used proportionally more resources than others. -Lea
@Hetzner_OL they do not even know what they use, how they use it or why do they need it. Some people (obviously) just have need to complain "everything Hetzner" lately.
So far, from what I could see they are either scammers, looks like,smells like, walks like scammer (but not 100% proved), pro-russians or russians that got suspended because of their own democracy and "right to choose" etc, etc..
So my advice (from someone that uses Hetzner for 10+ years without a glitch but also that spends less than 100 EUR per month).. just ignore it.
Thank you Lea but you did not answer the question in your quoted reply.
I miss Katie
Mostly time its not efficient on big scale even they doing it, disable
CONFIG_PARAVIRTyou can basically see on MSR address by checking on$ doas dmesg | grep kvm-
And there no such kvm-clock or kvm-stealtime, they probably disable it on hypervision.
My assumption only they doing CPU throttling on their hypervision to prevent abuse or indication underclock it on first place.
We*
I love Hetzner for their goofy responses (please don’t terminate my server I am begging)
I'm sure your customers are very pleased to have such a knowledgeable provider at their disposal. It's time to go back to virtualization 101.
Actual fact:
You raised your prices AND lowered the bandwidth by 20 TIMES, all for Black Friday.
Game over.
Lea, are you the one appearing in Hetzner YouTube shorts videos?
that escalated quickly
This is golden thread
We get it – change can be hard. That’s why we went out of our way to communicate the reasons for this adjustment on some of our social media platforms, which is more than most would do. So, we made sure to explain the reasons behind this change, as we thought transparency would help. Clearly, that wasn’t enough... But if you haven't seen our explanation for this price adjustment, just let us know so that we can also share it with the rest of you here.
We know that some of you were really hoping for those, and we're also sad that we didn't have any. -Lea
Hey y'all. You all are making me blush. It's nice to feel so wanted. I'm not sure what it is about me that you all miss so much, but it's giving me a bit of an ego-boost, which is always nice towards the end of the week.
We put out a statement last week about the price increase in the USA for the cloud products. Here it is again, in case you missed it:
"The good news is that there will be no pricing increases for our cloud locations in Europe! We are currently forced to pass on pricing increases from our colocation partner in the USA to our customers. This is the actual rationale behind the price adjustment. Additionally, we made the decision to create a fundamentally more equitable traffic pricing system when we launched our location in Singapore. The majority of our customers profit from a sliding scale of traffic expenses, according to our investigation of their traffic usage. In order to play with open cards, it was crucial that we make all of this known before launching the dollar as a payment option. Taking a step back, though, despite the price adjustment in the USA, we are playing in the top league in terms of price performance and will make every effort to continue doing so in the future."
I will be visiting family soon, so I will be away over the holidays. But perhaps I can shift around some of my other commitments next year so I can peek in at LET more often. --Katie
This is your FIFTH reply to the thread while not answering the main question.