@AnthonySmith said: I am interested in your opinion on something, there are some locations that are just expensive, no way is tierhive sustainable in those locations, do you think adding a multiplier, e.g. 1.5x token costs, obvious at the point of sale, is the right way to go or just don't bother with ultra exotic locations or something else?
Think Kenya, Lagos, iceland, Edinborough etc etc, not your standard locations.
Would love to see TierHive open up in Indonesia — Jakarta, Batam, Bogor, Surabaya would be awesome spots. Totally fine if you need to add a multiplier on the tokens, that’s fair.
@AnthonySmith said:
2. Storage VPS options now available in London AND Vint Hill USA, probably the cheapest option possible, if anyone knows cheaper, let me know!!
A little benchmarking with fio shows the HDDs(Vint Hill) running at 99/100 IOPS read/write. The root disk(Fast NVME) in Low IOPS mode clocks in at 55/57 IOPS read/write.
In this situation, moving the entire OS to the HDD may actually make the VPS disk IOPS ~100% faster, for an additional 0.000004 tokens/hr(for 1GB of HDD). No practical use perhaps, but it was a fun experiment nonetheless.
@AnthonySmith said: I am interested in your opinion on something, there are some locations that are just expensive, no way is tierhive sustainable in those locations, do you think adding a multiplier, e.g. 1.5x token costs, obvious at the point of sale, is the right way to go or just don't bother with ultra exotic locations or something else?
Think Kenya, Lagos, iceland, Edinborough etc etc, not your standard locations.
Would love to see TierHive open up in Indonesia — Jakarta, Batam, Bogor, Surabaya would be awesome spots. Totally fine if you need to add a multiplier on the tokens, that’s fair.
The plan is to be everywhere, naturally we are focusing on the cheaper locations first but if anyone has any recommendations on hosts in other locations I am keeping a list locally.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
@AnthonySmith said:
2. Storage VPS options now available in London AND Vint Hill USA, probably the cheapest option possible, if anyone knows cheaper, let me know!!
A little benchmarking with fio shows the HDDs(Vint Hill) running at 99/100 IOPS read/write. The root disk(Fast NVME) in Low IOPS mode clocks in at 55/57 IOPS read/write.
In this situation, moving the entire OS to the HDD may actually make the VPS disk IOPS ~100% faster, for an additional 0.000004 tokens/hr(for 1GB of HDD). No practical use perhaps, but it was a fun experiment nonetheless.
Hmm it's not quite that simple, you can't measure the iops that your VPS needs to exist in order to run the benchmark to begin with.
The hdd is not also running the os and also does not get the burstable IO that the main disk does.
Although that said running on low priority mode in most cases will cap you at around 200 iops currently on an active node.
Also, you can change the priority of the nvme unlike the HDD.
But I see what you mean either way, on the surface the numbers are odd.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
Edit: Forgot to say, IPv6 and HAProxy are also available, and Amsterdam has a 2.5 multiplier on network speed.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
@sshbox said:
What does network speed multiplier in AMS mean?
How is life of service defined?
What qualifies as an upgrade or downgrade?
P.S. You need to update your sig.
P.P.S Funky routing to Hetzner in Germany. Unexplained 20 ms hop.
It means VPS in Amsterdam at the moment instead of running at say 100mbit will run at 250mbit. If the tier is at 300mbit it will run at 750mbit
A change of tier on the network is an upgrade or downgrade.
Will update my signature when I am on a big screen because I am old.
At least it's not 40ms eh!
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
@sshbox said:
I don't see a knob to turn that specifies network bandwidth. Are network performance levels supposed to correlate with network bandwidth?
All in all I think your 2.5x promotion is confusing.
How is life of service defined? Will VPS stops/suspends affect eligibility?
Ok I will try to do better next time sorry.
I really struggle to explain the network tiers because people constantly want absolute numbers and that's not how it works.
If low is currently 50mbit and that is Equal to 1 or 100% then 2.5 is 250% or 150mbit, but the raw mbit number is fluid and based on the tier/priority you are paying for.
You have low medium and high tiers on the network that is your knob to turn.
I will get it better documented with examples.
The life of the service is the life of the service, if it exists you are paying for it, up down or suspended, when it no longer exists the life of the service of that VPS ends
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
Honestly network performance tiers are really confusing. I am not using network actively, but in my short tests there was virtually no difference between lowest and highest tiers.
My VPS in France rarely goes beyond 5-6mbit even in "high performance" mode, and it's around the same in "low performance". PL VPS does about 40mbit in both modes as well.
I don't ask for more bandwith or anything like that. It's just some sort of clarification what those tiers even mean should be helpful.
@Scioner said:
Honestly network performance tiers are really confusing. I am not using network actively, but in my short tests there was virtually no difference between lowest and highest tiers.
My VPS in France rarely goes beyond 5-6mbit even in "high performance" mode, and it's around the same in "low performance". PL VPS does about 40mbit in both modes as well.
I don't ask for more bandwith or anything like that. It's just some sort of clarification what those tiers even mean should be helpful.
Yeah I think it's a fair comment, I will have a think about it we might just have to set numbers and limits it's a constant source of confusion.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
@Scioner said:
Honestly network performance tiers are really confusing. I am not using network actively, but in my short tests there was virtually no difference between lowest and highest tiers.
My VPS in France rarely goes beyond 5-6mbit even in "high performance" mode, and it's around the same in "low performance". PL VPS does about 40mbit in both modes as well.
I don't ask for more bandwith or anything like that. It's just some sort of clarification what those tiers even mean should be helpful.
Yeah I think it's a fair comment, I will have a think about it we might just have to set numbers and limits it's a constant source of confusion.
To be a bit more clear - I do understand what tiers is some sort of scale for dynamic value, which is based on current capabilities. But having no difference after switching tiers is really confusing.
Like, yeah, my virtual number somewhere got bigger, that's probably good. But I don't see any changes, so why should I pay more.
@Scioner said:
Honestly network performance tiers are really confusing. I am not using network actively, but in my short tests there was virtually no difference between lowest and highest tiers.
My VPS in France rarely goes beyond 5-6mbit even in "high performance" mode, and it's around the same in "low performance". PL VPS does about 40mbit in both modes as well.
I don't ask for more bandwith or anything like that. It's just some sort of clarification what those tiers even mean should be helpful.
Yeah I think it's a fair comment, I will have a think about it we might just have to set numbers and limits it's a constant source of confusion.
To be a bit more clear - I do understand what tiers is some sort of scale for dynamic value, which is based on current capabilities. But having no difference after switching tiers is really confusing.
Like, yeah, my virtual number somewhere got bigger, that's probably good. But I don't see any changes, so why should I pay more.
I assume it's a priority thing. As the machine loads up you'd see more difference.
@Scioner said:
Honestly network performance tiers are really confusing. I am not using network actively, but in my short tests there was virtually no difference between lowest and highest tiers.
My VPS in France rarely goes beyond 5-6mbit even in "high performance" mode, and it's around the same in "low performance". PL VPS does about 40mbit in both modes as well.
I don't ask for more bandwith or anything like that. It's just some sort of clarification what those tiers even mean should be helpful.
Yeah I think it's a fair comment, I will have a think about it we might just have to set numbers and limits it's a constant source of confusion.
To be a bit more clear - I do understand what tiers is some sort of scale for dynamic value, which is based on current capabilities. But having no difference after switching tiers is really confusing.
Like, yeah, my virtual number somewhere got bigger, that's probably good. But I don't see any changes, so why should I pay more.
Yep it's a combination of perceptions, lack of visibility and understanding and maybe some backend issues I will check.
Right now under no contention even someone on low tier can use the full link speed if it's available, if someone on medium tier then needs bandwidth they will get a higher priority and same with high and medium.
There is a check and balance every 15 minutes and an automatic adjustment to ensure no one person can hig the link and to make sure everyone gets as much as possible.
While it was well intentioned it's confusing and probably just to messy for understanding as an end user.
The issue is we have different speeds in different locations at the Hypervisors so setting a simple limit is going to be rough.
I will give this some thought over the next few days, we might just have to do 10, 50 and 100 Mbit and be done with it.
I will let it stew in my head.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
@Scioner said:
Honestly network performance tiers are really confusing. I am not using network actively, but in my short tests there was virtually no difference between lowest and highest tiers.
My VPS in France rarely goes beyond 5-6mbit even in "high performance" mode, and it's around the same in "low performance". PL VPS does about 40mbit in both modes as well.
I don't ask for more bandwith or anything like that. It's just some sort of clarification what those tiers even mean should be helpful.
Yeah I think it's a fair comment, I will have a think about it we might just have to set numbers and limits it's a constant source of confusion.
To be a bit more clear - I do understand what tiers is some sort of scale for dynamic value, which is based on current capabilities. But having no difference after switching tiers is really confusing.
Like, yeah, my virtual number somewhere got bigger, that's probably good. But I don't see any changes, so why should I pay more.
I assume it's a priority thing. As the machine loads up you'd see more difference.
Yep this is the simple version.
But honestly I am getting worn down with ticket after ticket with people wanting numbers, it's my own fault really I have done a shit job of setting expectations.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
@Scioner said:
Honestly network performance tiers are really confusing. I am not using network actively, but in my short tests there was virtually no difference between lowest and highest tiers.
My VPS in France rarely goes beyond 5-6mbit even in "high performance" mode, and it's around the same in "low performance". PL VPS does about 40mbit in both modes as well.
I don't ask for more bandwith or anything like that. It's just some sort of clarification what those tiers even mean should be helpful.
Yeah I think it's a fair comment, I will have a think about it we might just have to set numbers and limits it's a constant source of confusion.
To be a bit more clear - I do understand what tiers is some sort of scale for dynamic value, which is based on current capabilities. But having no difference after switching tiers is really confusing.
Like, yeah, my virtual number somewhere got bigger, that's probably good. But I don't see any changes, so why should I pay more.
Yep it's a combination of perceptions, lack of visibility and understanding and maybe some backend issues I will check.
Right now under no contention even someone on low tier can use the full link speed if it's available, if someone on medium tier then needs bandwidth they will get a higher priority and same with high and medium.
There is a check and balance every 15 minutes and an automatic adjustment to ensure no one person can hig the link and to make sure everyone gets as much as possible.
While it was well intentioned it's confusing and probably just to messy for understanding as an end user.
The issue is we have different speeds in different locations at the Hypervisors so setting a simple limit is going to be rough.
I will give this some thought over the next few days, we might just have to do 10, 50 and 100 Mbit and be done with it.
I will let it stew in my head.
Not sure if it's a good idea, but maybe some sort of dynamic description on the instance and upgrade/downgrade screens would be sufficient? Like "under current load share of tier N is up to 5/10/20/90% of total bandwith". With "current" somehow emphasized.
It clears both confusions - why is my speed high on low tier, and what change would bring me stepping tier up (or down).
@Scioner said: It clears both confusions - why is my speed high on low tier, and what change would bring me stepping tier up (or down).
No, that's even worse, because people don't understand how numbers work. Availability does not equal guarantee unless you pay for it, even bursting. That's specifically why Ant is checking every 15 minutes.
"Hey, fucko- you're dragging everybody down" only works if they have enough funds/tokens to make them pay for it."
"It's a hard life- to be a stick insect." - Karl Pilkington
You and your colleague have done / are doing a great job, really!
To celebrate TierHive's new tenth location I added $10 to my account and received $11.00 because of the 10% bonus sale.. Probably will last nearly forever!
I I understand correctly, the speed issues discussed above might be vastly simplified by describing the 'Performance Tiers" as "priority when there is contention." When there is no contention, everyone gets the maximum.
Additionally, you could state the instance's "Base Speed." Then factors like 2.5 x base speed are clear.
@AnthonySmith said:
we have different speeds in different locations at the Hypervisors
Perhaps describing the "Performance Tiers" and listing the base speeds somehow might be integrated into the VPS creation page? Someone who wants to maximize speed then easily could select a location with high base speed.
I continue to wish TierHive good luck and amazing progress!
Maybe we just give you xxx Mbit and make that a global cap.
So for example 0.000005 per Mbit burst per hour with a maximum.
You can then allocate Mbit /s to your VPS.
1 Mbit x 250 VPS or 250Mbit to 1 VPS
Idk, I will come up with a bigger list of options including doing nothing but being more transparent, and ask for input after I check feasibility.
It's in the multiple tickets per day realm now so something is obviously needed.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
They're just trying to find some simple, if morally unethical way to fuck you. Block all third world countries and be done with it. Doubly so if they're getting free service from Google, Microsoft, et al.
"It's a hard life- to be a stick insect." - Karl Pilkington
@WSS said:
They're just trying to find some simple, if morally unethical way to fuck you. Block all third world countries and be done with it. Doubly so if they're getting free service from Google, Microsoft, et al.
So... just block India?
While I dont support what WSS says, I do believe you should consider tracking metrics of which country people are opening more support request, what are the request about and build a dynamic FAQ from that.
I speak fluent sarcasm and broken logic. | I would agree with you, but thæn we’d both be wrong.
This is what we are going to start doing; it might not be right for every circumstance, but this feels the most complete way of doing things that fits what we already have without massive rewrites.
This will be documented, but this is while it's fresh in my head. Ignore the specific speed numbers below; they are just examples.
Network tiers will start to show the actual speed limits; we have just set arbitrary speeds so people know what they are getting in advance.
Every tier will get a 100% burst for 10 seconds, so if you have 50 Mbps, you can actually do 100 Mbps for 10 seconds.
Every location will show its specific speeds, so in Singapore, low might be say 5 Mbps while in France, low will be 20 Mbps.
At a per location basis your total traffic on the bridge will be tracked, and we are going to set a min/max 95th percentile figure per location, which will be published.
If the min is 10 Mbps and the max is 100 Mbps and you hit 100 Mbps on average for 1 full hour, the bridge will be throttled to 90mbit if you then hit 90mbit for the next hour on average, you will be throttled to 80mbit and so on right down to 10mbit, if you are under the figure then you go back up 10mbit per hour until the limit is removed entirley.
What I have observed this morning and what the past data shows is that about 5% of the users use about 95% of the available links, and our backend is doing what it should do but they just keep spinning up more and more VPS to spread it out so the issue needs to be dealt with at a bridge interface level, which we can do as we give each customer their own bridge.
The chances of any legitimate/suitable TierHive use hiting 100Mbit 24x7 is 0, if you need that its already the wrong service so based on the data available, this change will actually impact about 5% of customers in reality, while restoring better speeds to 95% in some regions.
The changes from outside are actually pretty minimal its just 1 extra table to know what and if you are limited and we already capture the traffic data we need and we just need to adapt our current QOS system to target the bridges instead of the individual VPS so this has already started happening. It will be pushed into production over the weekend, hopefully.
This way if you abuse the unmetered traffic with 300mbit constantly from 1 VPS or 3mbit from 100 VPS you will be gradually limited 10mbit at a time on a per-location basis instead of a per-VPS basis, and it self relaxes 10mbit at a time.
We will tweak the numbers until we find the right balance.
Hopefully that makes sense and will bring some clarity to the UI.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
All I ask is that the changes are communicated to customers and current limits are show in the VPS dashboard.
I was very annoyed when my VPS hit the (secret) CPU limit and I wasn't told about it. All I could see was that steal was 50-75%.
Is there a reason why CPU limits cannot be dynamically applied like the network shaping? I'm still a bit miffed about having to had to reboot my VPS to clear the CPU limiters.
For the record, I am not a heavy CPU user nor do I plan to be. I just happened to do some CPU heavy stuff during setup and testing of my first VPS.
actual network tiers shown by Mbps now on the vps creation page and on the upgrade / downgrade tab on the vps management page.
Sadly this has meant we have had to fix speeds per tier instead of having them dynamically shift but I have increased the burst duration to compensate to some degree.
We are not going to start enforcement until the weekend, that part is still work in progress.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
@sshbox said:
All I ask is that the changes are communicated to customers and current limits are show in the VPS dashboard.
I was very annoyed when my VPS hit the (secret) CPU limit and I wasn't told about it. All I could see was that steal was 50-75%.
Is there a reason why CPU limits cannot be dynamically applied like the network shaping? I'm still a bit miffed about having to had to reboot my VPS to clear the CPU limiters.
For the record, I am not a heavy CPU user nor do I plan to be. I just happened to do some CPU heavy stuff during setup and testing of my first VPS.
I just communicated it.
Honestly, if you get annoyed with TierHive... dont use it, its not the right product for you, sorry if that sounds harsh but you obviously missed the point.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL FREE tokens on sign up, try before you buy. | Join us on Reddit
Please Anthony, I am not trying to annoy you, and I ask this with only love in my heart: what is the point?
Have you considered that perhaps the point was not communicated clearly and expectations weren't set properly at the beginning?
To quote your own web copy:
"Hourly VPS hosting that scales with you ... You just do what you do, let us worry about the rest."
I know that the large print giveth and the small print taketh away, but surely your front page could also offer some guidance on acceptable use and limits? A naïve user will go that sounds great, select all high priority on all tier, figure they are all set and go to town. This (obviously in hindsight) is not the intended use, but things that are obvious to you are not obvious to everyone.
I've had the (not so) great fortune of having to deal with the unwashed masses in a lot of B2C services and it really helps to spell out the obvious.
In any case, please don't take this too hard, these are just my opinions, and we all have plenty of those to go around. Cheers, mate.
Comments
Would love to see TierHive open up in Indonesia — Jakarta, Batam, Bogor, Surabaya would be awesome spots. Totally fine if you need to add a multiplier on the tokens, that’s fair.
A little benchmarking with fio shows the HDDs(Vint Hill) running at 99/100 IOPS read/write. The root disk(Fast NVME) in Low IOPS mode clocks in at 55/57 IOPS read/write.
In this situation, moving the entire OS to the HDD may actually make the VPS disk IOPS ~100% faster, for an additional 0.000004 tokens/hr(for 1GB of HDD). No practical use perhaps, but it was a fun experiment nonetheless.
The plan is to be everywhere, naturally we are focusing on the cheaper locations first but if anyone has any recommendations on hosts in other locations I am keeping a list locally.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
Hmm it's not quite that simple, you can't measure the iops that your VPS needs to exist in order to run the benchmark to begin with.
The hdd is not also running the os and also does not get the burstable IO that the main disk does.
Although that said running on low priority mode in most cases will cap you at around 200 iops currently on an active node.
Also, you can change the priority of the nvme unlike the HDD.
But I see what you mean either way, on the surface the numbers are odd.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
Amsterdam is live
Edit: Forgot to say, IPv6 and HAProxy are also available, and Amsterdam has a 2.5 multiplier on network speed.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
What does network speed multiplier in AMS mean?
How is life of service defined?
What qualifies as an upgrade or downgrade?
P.S. You need to update your sig.
P.P.S Funky routing to Hetzner in Germany. Unexplained 20 ms hop.
It means VPS in Amsterdam at the moment instead of running at say 100mbit will run at 250mbit. If the tier is at 300mbit it will run at 750mbit
A change of tier on the network is an upgrade or downgrade.
Will update my signature when I am on a big screen because I am old.
At least it's not 40ms eh!
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
I don't see a knob to turn that specifies network bandwidth. Are network performance levels supposed to correlate with network bandwidth?
All in all I think your 2.5x promotion is confusing.
How is life of service defined? Will VPS stops/suspends affect eligibility?
I got this:
Yes
OK
No
(I think)
PS Sorry - I've been drinking.
Ok I will try to do better next time sorry.
I really struggle to explain the network tiers because people constantly want absolute numbers and that's not how it works.
If low is currently 50mbit and that is Equal to 1 or 100% then 2.5 is 250% or 150mbit, but the raw mbit number is fluid and based on the tier/priority you are paying for.
You have low medium and high tiers on the network that is your knob to turn.
I will get it better documented with examples.
The life of the service is the life of the service, if it exists you are paying for it, up down or suspended, when it no longer exists the life of the service of that VPS ends
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
Honestly network performance tiers are really confusing. I am not using network actively, but in my short tests there was virtually no difference between lowest and highest tiers.
My VPS in France rarely goes beyond 5-6mbit even in "high performance" mode, and it's around the same in "low performance". PL VPS does about 40mbit in both modes as well.
I don't ask for more bandwith or anything like that. It's just some sort of clarification what those tiers even mean should be helpful.
Yeah I think it's a fair comment, I will have a think about it we might just have to set numbers and limits it's a constant source of confusion.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
To be a bit more clear - I do understand what tiers is some sort of scale for dynamic value, which is based on current capabilities. But having no difference after switching tiers is really confusing.
Like, yeah, my virtual number somewhere got bigger, that's probably good. But I don't see any changes, so why should I pay more.
I assume it's a priority thing. As the machine loads up you'd see more difference.
Yep it's a combination of perceptions, lack of visibility and understanding and maybe some backend issues I will check.
Right now under no contention even someone on low tier can use the full link speed if it's available, if someone on medium tier then needs bandwidth they will get a higher priority and same with high and medium.
There is a check and balance every 15 minutes and an automatic adjustment to ensure no one person can hig the link and to make sure everyone gets as much as possible.
While it was well intentioned it's confusing and probably just to messy for understanding as an end user.
The issue is we have different speeds in different locations at the Hypervisors so setting a simple limit is going to be rough.
I will give this some thought over the next few days, we might just have to do 10, 50 and 100 Mbit and be done with it.
I will let it stew in my head.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
Yep this is the simple version.
But honestly I am getting worn down with ticket after ticket with people wanting numbers, it's my own fault really I have done a shit job of setting expectations.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
Not sure if it's a good idea, but maybe some sort of dynamic description on the instance and upgrade/downgrade screens would be sufficient? Like "under current load share of tier N is up to 5/10/20/90% of total bandwith". With "current" somehow emphasized.
It clears both confusions - why is my speed high on low tier, and what change would bring me stepping tier up (or down).
No, that's even worse, because people don't understand how numbers work. Availability does not equal guarantee unless you pay for it, even bursting. That's specifically why Ant is checking every 15 minutes.
"Hey, fucko- you're dragging everybody down" only works if they have enough funds/tokens to make them pay for it."
"It's a hard life- to be a stick insect." - Karl Pilkington
You and your colleague have done / are doing a great job, really!
To celebrate TierHive's new tenth location I added $10 to my account and received $11.00 because of the 10% bonus sale.. Probably will last nearly forever!
I I understand correctly, the speed issues discussed above might be vastly simplified by describing the 'Performance Tiers" as "priority when there is contention." When there is no contention, everyone gets the maximum.
Additionally, you could state the instance's "Base Speed." Then factors like 2.5 x base speed are clear.
Perhaps describing the "Performance Tiers" and listing the base speeds somehow might be integrated into the VPS creation page? Someone who wants to maximize speed then easily could select a location with high base speed.
I continue to wish TierHive good luck and amazing progress!
Maybe we just give you xxx Mbit and make that a global cap.
So for example 0.000005 per Mbit burst per hour with a maximum.
You can then allocate Mbit /s to your VPS.
1 Mbit x 250 VPS or 250Mbit to 1 VPS
Idk, I will come up with a bigger list of options including doing nothing but being more transparent, and ask for input after I check feasibility.
It's in the multiple tickets per day realm now so something is obviously needed.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
FAQ in the knowledge base. Check box "I have read the FAQ" before being able to create a ticket.
Lock everyone to 56k. If you can send Hayes commands, it gets doubled to 128k.
"It's a hard life- to be a stick insect." - Karl Pilkington
They're just trying to find some simple, if morally unethical way to fuck you. Block all third world countries and be done with it. Doubly so if they're getting free service from Google, Microsoft, et al.
"It's a hard life- to be a stick insect." - Karl Pilkington
So... just block India?
While I dont support what WSS says, I do believe you should consider tracking metrics of which country people are opening more support request, what are the request about and build a dynamic FAQ from that.
Perhaps it's a language thing. English is my first language and I had no problem understanding. Don't be too hard on yourself.
Keith
ok so we had a chat.
This is what we are going to start doing; it might not be right for every circumstance, but this feels the most complete way of doing things that fits what we already have without massive rewrites.
This will be documented, but this is while it's fresh in my head. Ignore the specific speed numbers below; they are just examples.
What I have observed this morning and what the past data shows is that about 5% of the users use about 95% of the available links, and our backend is doing what it should do but they just keep spinning up more and more VPS to spread it out so the issue needs to be dealt with at a bridge interface level, which we can do as we give each customer their own bridge.
The chances of any legitimate/suitable TierHive use hiting 100Mbit 24x7 is 0, if you need that its already the wrong service so based on the data available, this change will actually impact about 5% of customers in reality, while restoring better speeds to 95% in some regions.
The changes from outside are actually pretty minimal its just 1 extra table to know what and if you are limited and we already capture the traffic data we need and we just need to adapt our current QOS system to target the bridges instead of the individual VPS so this has already started happening. It will be pushed into production over the weekend, hopefully.
This way if you abuse the unmetered traffic with 300mbit constantly from 1 VPS or 3mbit from 100 VPS you will be gradually limited 10mbit at a time on a per-location basis instead of a per-VPS basis, and it self relaxes 10mbit at a time.
We will tweak the numbers until we find the right balance.
Hopefully that makes sense and will bring some clarity to the UI.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
All I ask is that the changes are communicated to customers and current limits are show in the VPS dashboard.
I was very annoyed when my VPS hit the (secret) CPU limit and I wasn't told about it. All I could see was that steal was 50-75%.
Is there a reason why CPU limits cannot be dynamically applied like the network shaping? I'm still a bit miffed about having to had to reboot my VPS to clear the CPU limiters.
For the record, I am not a heavy CPU user nor do I plan to be. I just happened to do some CPU heavy stuff during setup and testing of my first VPS.
actual network tiers shown by Mbps now on the vps creation page and on the upgrade / downgrade tab on the vps management page.
Sadly this has meant we have had to fix speeds per tier instead of having them dynamically shift but I have increased the burst duration to compensate to some degree.
We are not going to start enforcement until the weekend, that part is still work in progress.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
I just communicated it.
Honestly, if you get annoyed with TierHive... dont use it, its not the right product for you, sorry if that sounds harsh but you obviously missed the point.
TierHive - Hourly VPS - NAT Native - /24 per customer - DE, UK, SG, CA, USA x3, FR, AU, PL, NL
FREE tokens on sign up, try before you buy. | Join us on Reddit
I don't think all customers religiously frequent LES. Anyways, please also show current limits on the dashboard.
Please Anthony, I am not trying to annoy you, and I ask this with only love in my heart: what is the point?
Have you considered that perhaps the point was not communicated clearly and expectations weren't set properly at the beginning?
To quote your own web copy:
"Hourly VPS hosting that scales with you ... You just do what you do, let us worry about the rest."
I know that the large print giveth and the small print taketh away, but surely your front page could also offer some guidance on acceptable use and limits? A naïve user will go that sounds great, select all high priority on all tier, figure they are all set and go to town. This (obviously in hindsight) is not the intended use, but things that are obvious to you are not obvious to everyone.
I've had the (not so) great fortune of having to deal with the unwashed masses in a lot of B2C services and it really helps to spell out the obvious.
In any case, please don't take this too hard, these are just my opinions, and we all have plenty of those to go around. Cheers, mate.