3 ms·

Vultr's Bandwidth Policy “Improvements”

Just received email from Vultr: they suddenly re-announce "improvements" in their bandwidth policy. One thing they've added on top of November's announcement[1] is that they are removing (!) "Stop Servers on Bandwidth Usage" functionality. I suspect that a good chunk of some of VPS providers' profits are accidental overages[2] (and then those providers sometimes get "aww" from the crowd by forgiving /some/ of the charges, which were totally avoidable to begin with). This kind of practice is despicable, IMO. Any recommendations from HN on VPS providers that never overcharge? [1] https://www.vultr.com/news/Vultr-Announces-Reduced-Bandwidth-Pricing-2-Tb-Of-Free-Monthly-Egress-Free-Ingress-And-Global-Pooling/ [2] https://news.ycombinator.com/item?id=30054739 https://www.digitalocean.com/community/questions/droplet-bandwidth-overage-ridiculous-amount ------------------ Subject: Vultr.com - Notification of bandwidth policy improvements and changes to related functionality. In November, we announced updates to our bandwidth billing policy, which will make high-performance cloud computing even more accessible and affordable for users around the world. You can view the announcement here. Here are the highlights: - Every Vultr customer will now receive two terabytes (2 TB) of free outbound data transfer every month, in addition to the transfer included with each subscription. - Data ingress will now be free. - All customers will receive global account-level pooling of transfer across all instances and locations. - Vultr will now offer a reduced, single worldwide egress overage rate of only $0.01 per GB. These improvements also some of the existing functionality. We will be making the following changes: - The "Stop Servers on Bandwidth Usage" function will be discontinued. - The "Bandwidth Usage Notifications" will be updated to reflect the new policy and will be calculated based on the pooled bandwidth total instead of on a per-instance basis. If you have any questions or concerns about this change, reach out to our customer support team for assistance. We value your patronage and are dedicated to providing the highest level of service. ----------------------

LinuxBender 4y ago
I don't have an answer to which VPS providers still have the feature to disable a VM at a certain bw usage. The only thing I can add is that this behavior can be accomplished on any node running Linux by using one of the bandwidth related modules such as quota. [1] One advantage of using IPTables to do this is that one can enforce it on any or multiple ports [2] but exclude your management ports such as SSHD. The next matching rule could be a tcp-reset only responding to your automation so that it knows to take that node out of any pools one has defined. [1] - https://ipset.netfilter.org/iptables-extensions.man.html#lbBT https://ipset.netfilter.org/iptables-extensions.man.html#lbB... [2] - https://ipset.netfilter.org/iptables-extensions.man.html#lbBM https://ipset.netfilter.org/iptables-extensions.man.html#lbB...
- PetroulaNelopi 4y ago
  Thanks for the suggestion! The problem with self-imposed quotas comes up when provider and customer disagrees on the charges. Provider may simply disagree with you and insist that you did in fact used "X" amount more of bandwidth. This "X" can be small, or astronomical, it all depends on how trustworthy the provider is, or how good your skills of setting up the quotas are (in case of own mistakes.) Now, given that the provider already tries to pull shady stuff on me -- how much trust should I have in case of a dispute?.. I'd rather prefer to avoid the whole possibility of such dispute, frankly.
  - LinuxBender 4y ago
    I should have stayed logged into Vultr. I believe they like all the others have bandwidth charts that show usage per VM and per account. Not all of them pool the allotted bandwidth. The same systems that build those graphs are the ones controlling bw over-usage. Their graphs should line up with network interface statistics and iptables counters minus their internal management traffic. In my opinion if they are doing something shady with the bandwidth graphs that should be brought to the attention of the internet ASAP and I would expect their behavior to improve unless their goal is to file chapter 11. This should apply to all VPS providers. Have you seen evidence of Vultr or any other VPS providers fudging the bandwidth graphs? If you suspect this is happening the easiest thing to do is create a daily cron job that grabs all the interface statistics and iptables counters to a text file daily and then rsync those to a centralized system. Or if you prefer having your own graphs some people use Munin [1] or any of the alternative options [2]. [1] - https://munin-monitoring.org/ https://munin-monitoring.org/ [2] - https://alternativeto.net/software/munin/ https://alternativeto.net/software/munin/
msh 4y ago
The only reason I stick with scaleway. No direct data charges.

LinuxBender 4y ago

I don't have an answer to which VPS providers still have the feature to disable a VM at a certain bw usage. The only thing I can add is that this behavior can be accomplished on any node running Linux by using one of the bandwidth related modules such as quota. [1] One advantage of using IPTables to do this is that one can enforce it on any or multiple ports [2] but exclude your management ports such as SSHD. The next matching rule could be a tcp-reset only responding to your automation so that it knows to take that node out of any pools one has defined. [1] - https://ipset.netfilter.org/iptables-extensions.man.html#lbBT https://ipset.netfilter.org/iptables-extensions.man.html#lbB... [2] - https://ipset.netfilter.org/iptables-extensions.man.html#lbBM https://ipset.netfilter.org/iptables-extensions.man.html#lbB...

PetroulaNelopi 4y ago
Thanks for the suggestion! The problem with self-imposed quotas comes up when provider and customer disagrees on the charges. Provider may simply disagree with you and insist that you did in fact used "X" amount more of bandwidth. This "X" can be small, or astronomical, it all depends on how trustworthy the provider is, or how good your skills of setting up the quotas are (in case of own mistakes.) Now, given that the provider already tries to pull shady stuff on me -- how much trust should I have in case of a dispute?.. I'd rather prefer to avoid the whole possibility of such dispute, frankly.
- LinuxBender 4y ago
  I should have stayed logged into Vultr. I believe they like all the others have bandwidth charts that show usage per VM and per account. Not all of them pool the allotted bandwidth. The same systems that build those graphs are the ones controlling bw over-usage. Their graphs should line up with network interface statistics and iptables counters minus their internal management traffic. In my opinion if they are doing something shady with the bandwidth graphs that should be brought to the attention of the internet ASAP and I would expect their behavior to improve unless their goal is to file chapter 11. This should apply to all VPS providers. Have you seen evidence of Vultr or any other VPS providers fudging the bandwidth graphs? If you suspect this is happening the easiest thing to do is create a daily cron job that grabs all the interface statistics and iptables counters to a text file daily and then rsync those to a centralized system. Or if you prefer having your own graphs some people use Munin [1] or any of the alternative options [2]. [1] - https://munin-monitoring.org/ https://munin-monitoring.org/ [2] - https://alternativeto.net/software/munin/ https://alternativeto.net/software/munin/

msh 4y ago

The only reason I stick with scaleway. No direct data charges.