Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


High uptime VPS hosting
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.

High uptime VPS hosting

treesmokahtreesmokah Member
edited February 2023 in General

Hey,

I'm in need for a high uptime VPS box for server-monitoring purposes.
Very little bandwidth and about 2 vCPU / 2 or 4 GB of ram per box.
EU or Scandinavian location. Budget is very flexible, and I'm willing to pay "premium" prices for a good product.

I have checked Vultr, Linode and other "high-end" vps providers and they do indeed look nice.
Do you guys have any experience with them?

I'm also open to suggestions based on Xen with multi-cluster(or preferably, geo diverse) setup.

Thanks

Thanked by 1dusst
«1

Comments

  • ericlsericls Member, Patron Provider

    Have you considered doing fallbacks between different nodes?

  • @ericls said:
    Have you considered doing fallbacks between different nodes?

    It is an option for sure, from what I recall - Vultr supports doing "snapshots" but I'm unsure whether they can be exported or not, and if they cant - it sucks ass, if their storage clusters go bust.

  • My experiences for high uptime requirements are based on running tendermint nodes for validating proof of stake based crypto.

    Node needs to process a new block at around every 5 seconds depending on network. While it's normal to miss some blocks every day, a node on a good server can run without missing any blocks.

    I've used and tested a lot of low end VPSs for this purpose but I've ended up using Vultr high frequency instances in their UK location. It has been extremely solid, basically running months straight with perfect record for processing blocks.

    It is currently the only provider I truly trust to stay online and functional (without making a cluster of multiple VPS instances). I'm however moving towards a Docker Swarm based cluster solution which will allow even more certainty by having an automated fail over.

    Thanked by 2treesmokah maverick
  • @covent said:
    My experiences for high uptime requirements are based on running tendermint nodes for validating proof of stake based crypto.

    Node needs to process a new block at around every 5 seconds depending on network. While it's normal to miss some blocks every day, a node on a good server can run without missing any blocks.

    I've used and tested a lot of low end VPSs for this purpose but I've ended up using Vultr high frequency instances in their UK location. It has been extremely solid, basically running months straight with perfect record for processing blocks.

    It is currently the only provider I truly trust to stay online and functional (without making a cluster of multiple VPS instances). I'm however moving towards a Docker Swarm based cluster solution which will allow even more certainty by having an automated fail over.

    Thank you very much for sharing your experience.

    I was also thinking about "managed" Docker containers/Kubernetes - but god, I fucking hate Docker.
    Are there any providers doing IaaS LXC/LXD like with Kubernetes? Failover and all that good stuff.

  • conceptconcept Member
    edited February 2023

    I've always just used Hetzner Cloud for high uptime and it has worked really well for me. Occasionally, I would use Vultr or Azure. I've never really delved deep into Kubernetes stuff but I know DigitalOcean/Vultr makes things pretty easy to setup.

    Thanked by 2treesmokah nick_
  • @concept said:
    I've always just used Hetzner Cloud for high uptime and it has worked really well for me. Occasionally, I would use Vultr or Azure. I've never really delved deep into Kubernetes stuff but I know DigitalOcean/Vultr makes things pretty easy to setup.

    Thank you very much for sharing your experience.

    Sadly Hetzner is a no-go for me, I hate them.

    Thanked by 1concept
  • I would personally go with vultr or the big cloud giants.

    Thanked by 1treesmokah
  • My only production VPS for at least 6 months now is a Vultr High Frequency VPS in Stockholm.

    It serves websites literally paying the salary for several people, and I feel very good about my choice.

    Thanked by 1treesmokah
  • aquaaqua Member, Patron Provider

    If you don't care about your pockets hurting, go for the leading hosting company, AWS. Never had issues with our AWS servers, and I doubt I ever will.

  • @aqua said:
    If you don't care about your pockets hurting, go for the leading hosting company, AWS. Never had issues with our AWS servers, and I doubt I ever will.

    I may sound like I have issues with everyone lol but I won't put a single dollar in Amazon's pockets.
    I want Amazon to burn.

    Thanks for the suggestion though.

    Thanked by 1darkimmortal
  • treesmokahtreesmokah Member
    edited February 2023

    I do not mind a smaller provider, as long as they can provide good SLA(99.8+) and good support at similar prices to Vultr. At least 2 EU locations is a must though.

  • I've recently been migrating mainly to dedicated servers (although the transition has been slow for various reasons), but I still have a variety of VPS instances at different providers, "premium" and "non-premium" and several in between. I'm slowly phasing them out though to stick with dedicated when possible.

    Anyway, yes, the other "premium" services listed in this thread have good uptime too. The best uptime I've experienced for VPS services overall include: Vultr, Linode, Digital Ocean, AWS (I know you hate them), and Hetzner Cloud. Not necessarily in that order. But uptime has been consistently 99.95+% for most of those in my case over several years, not including user error on my end. :# Several of them 99.99+% uptime, and a couple of them full 100% uptime, depending on data center and maybe luck. I had instances at Vultr and Hetzner Cloud that were both 100% though. In fact, the more I think about it, I think all the HF instances at Vultr were 100% TBH. Maybe I was very lucky.

    All of that is anecdotal though. As are everyone's experiences here. It's too small of a sample group to know the real story. And YMMV since you have different needs and you'll be in different data centers, use a different stack, have different luck with what nodes you are on, etc...

    So if you need two for your project, I'd suggest getting one each from two different providers. Split the risk, get your own data points over time for your unique situation. Build your own trust relationship with each provider and your specific stack.

    Good luck.

  • Oh, I forgot to mention one other provider that I recall had great uptime -- Dreamhost's "DreamCompute" cloud service. In my experience, that was also very solid. However, I really didn't like their provisioning system so I phased them out a while back. But they did have great uptime. However, I believe the DreamCompute datacenter is in Virginia, so that won't qualify for your needs.

  • treesmokahtreesmokah Member
    edited February 2023

    @jlet88 said:
    I've recently been migrating mainly to dedicated servers (although the transition has been slow for various reasons), but I still have a variety of VPS instances at different providers, "premium" and "non-premium" and several in between. I'm slowly phasing them out though to stick with dedicated when possible.

    Anyway, yes, the other "premium" services listed in this thread have good uptime too. The best uptime I've experienced for VPS services overall include: Vultr, Linode, Digital Ocean, AWS (I know you hate them), and Hetzner Cloud. Not necessarily in that order. But uptime has been consistently 99.95+% for most of those in my case over several years, not including user error on my end. :# Several of them 99.99+% uptime, and a couple of them full 100% uptime, depending on data center and maybe luck. I had instances at Vultr and Hetzner Cloud that were both 100% though. In fact, the more I think about it, I think all the HF instances at Vultr were 100% TBH. Maybe I was very lucky.

    All of that is anecdotal though. As are everyone's experiences here. It's too small of a sample group to know the real story. And YMMV since you have different needs and you'll be in different data centers, use a different stack, have different luck with what nodes you are on, etc...

    So if you need two for your project, I'd suggest getting one each from two different providers. Split the risk, get your own data points over time for your unique situation. Build your own trust relationship with each provider and your specific stack.

    Good luck.

    Thank you very much for sharing your experience.

    The downside of dedicated servers is that they can obviously break.
    VPS can be migrated across nodes and have unmatched uptime compared to dedis.

    I mainly need a high uptime vps for monitoring my own infrastructure(own colocated hardware etc).
    I have monitoring tools in house, however I want to have something sitting in a different datacenter and network just to make sure everything is looking good from the "outside".

    I doubt I will be a "recognizable" client to gigants like Vultr because I only want to use them for headless monitoring with no server privileges.
    I want to keep everything as secure as possible and make sure that only I have access to core parts of my infrastructure(full disk encryption on all servers and all that good stuff).

    @jlet88 said:
    Oh, I forgot to mention one other provider that I recall had great uptime -- Dreamhost's "DreamCompute" cloud service. In my experience, that was also very solid. However, I really didn't like their provisioning system so I phased them out a while back. But they did have great uptime. However, I believe the DreamCompute datacenter is in Virginia, so that won't qualify for your needs.

    Yes, I know who Dreamhost is and they are really solid and in my eyes ethical(they support net-neutrality and free speech).
    Sadly for this usecase, the location is important to me - particularly routing and local laws(even though this box won't have any "special" access to my infra) - I do not want CIA goons lurking through my infra stats.

  • So your goals is to find a solid provider where you can get 2 vps's in different locations?

    Gcore is solid

  • @treesmokah said:
    The downside of dedicated servers is that they can obviously break.
    VPS can be migrated across nodes and have unmatched uptime compared to dedis.

    Very true. I mainly mentioned dedicated servers for unneeded background info :D not really recommending them for your situation.

    But yes, I get what you want to do. I think you have a great list of potential premium providers to work with, but you're only going to get a very limited sample set of data from this (or any) forum. I'd suggest diving in to the ones with the best reputations and maybe have 2-3 different additional providers until you establish a strong confidence level. The best data points are you own, right?

    In my case, I never would have thought Hetzner Cloud was going to be as reliable as it has been for the projects I've used it on. But I only gained confidence in it after years of use. Hetzner is inconsistent though, as I had a bad experience with their storage box service. But their cloud has been surprisingly good. Again, YMMV.

    And BTW, I can't speak for Microsoft Azure and Google Cloud Platform... I prefer not to use those services, sort of like how you prefer not to use AWS (although I wouldn't be as vehement in my desire for their demise). But Azure and Google have great uptime records too. And I have not touched Oracle yet, and probably never will, but I know some people love Oracle's cloud.

    Good luck.

  • bgerardbgerard Member
    edited February 2023

    ^^ Azure is pretty solid in terms of machine uptime. Not had an issue there. There was a network blip the other day but you'll probably find the experience is similar with all 3 of the major players.

    Obviously nothing is 100%, I'd go with two machines split between Azure, AWS, GCP, Linode or Vultr. Used go be a DO fan but it has been a while.

    Alternatively, could you use FaaS and a cron trigger for your purposes?

  • Hetzner has always had an excellent uptime for me

  • @kait said:
    So your goals is to find a solid provider where you can get 2 vps's in different locations?

    Gcore is solid

    I've been using Gcore for DNS for quite some time now, no issues.
    Time to give their VPS offering a shot ig.

    Thanked by 1kait
  • @jlet88 said: But Azure and Google have great uptime records too. And I have not touched Oracle yet, and probably never will, but I know some people love Oracle's cloud.

    I could give Google a shot ig, Azure and especially Oracle are a no-go.
    Anything Oracle or M$ touches turns in to literal dogshit.

  • ATeamATeam Member

    I had experienced with Vultr, I must admit their server fee is relative cheap, but sometimes I had issue accessing the server or it can automatically kicked me out from the server. When I asked this issue with their support team, they said 'there is no issue with the server/network'. Hmm...

    Maybe you may check Azure, but they are quite expensive. Or Hostforlife.eu can be good consideration for EU location.

    Thanked by 1treesmokah
  • OVH cloud instance with gravelines location, never had downtime for me. You can spin up a discovery model instance with 13$ for 4 GB Ram and 2 vCPU with 250 MBPS unmetered.

  • @sreekanth850 said:
    OVH cloud instance with gravelines location, never had downtime for me. You can spin up a discovery model instance with 13$ for 4 GB Ram and 2 vCPU with 250 MBPS unmetered.

    I want to avoid OVH, support is non existent and I had terrible experience with them.

  • WebProjectWebProject Host Rep, Veteran

    @treesmokah said:

    @sreekanth850 said:
    OVH cloud instance with gravelines location, never had downtime for me. You can spin up a discovery model instance with 13$ for 4 GB Ram and 2 vCPU with 250 MBPS unmetered.

    I want to avoid OVH, support is non existent and I had terrible experience with them.

    The support required additional steps in order to push the issues, otherwise if you know how to manage VMs / dedicated server so you don’t need any their support.

  • treesmokahtreesmokah Member
    edited March 2023

    @WebProject said:

    @treesmokah said:

    @sreekanth850 said:
    OVH cloud instance with gravelines location, never had downtime for me. You can spin up a discovery model instance with 13$ for 4 GB Ram and 2 vCPU with 250 MBPS unmetered.

    I want to avoid OVH, support is non existent and I had terrible experience with them.

    The support required additional steps in order to push the issues, otherwise if you know how to manage VMs / dedicated server so you don’t need any their support.

    I do know how to mange my services, but their own panel and billing issues are out of my reach.
    Been locked from my account for over a week because they had some funny issue.

    I prefer talking to real humans not some robots.
    Had to re-create a ticket 2 times because first person did not understand what I meant and I was not able to reply to a fucking ticket due to "insufficient permissions" even though I was using my main account to manage everything.

    So yeah, fuck them. I will never use them for production.

    Even deleting my company account was a hassle because of ticket system being all over the place and them taking at least a week to respond with a scripted message that had nothing to do with what I said.

    OVH is a McDonald's of hosting, some people like it - but anyone with higher expectations than "it works" will be pissed.

    Thanked by 1maverick
  • sreekanth850sreekanth850 Member
    edited March 2023

    @treesmokah said:

    @sreekanth850 said:
    OVH cloud instance with gravelines location, never had downtime for me. You can spin up a discovery model instance with 13$ for 4 GB Ram and 2 vCPU with 250 MBPS unmetered.

    I want to avoid OVH, support is non existent and I had terrible experience with them.

    If you need support you have to subscribe to their professional support packages/plans. Without that they will only respond to hardware related issues. that too 3 to 4 response per day. Too Slow.

    Using their cloud instances for last 1.5 years without any support. They have one of the best DDoS protection.

  • treesmokahtreesmokah Member
    edited March 2023

    @sreekanth850 said: If you need support you have to subscribe to their professional support packages/plans.

    I won't subscribe to a support plan and pay hundreds per month for an unmanaged service that normally costs me $5.
    I do not need management, I want them to fix their Alibaba-like panel.
    Not to mention their website is split to 2858258025205 regions and even if you register on their "International" one - You won't find certain countries on the list for whatever reason(and there is no option to switch the region nor create a ticket telling them to do it for you - there is no email for contact only their ticket panel which is accessible ONLY after filling personal details which is not possible due to country not being on the list). It is a fucking joke.

    Purchased a VPS by mistake or something along these lines? Forget about a refund(which is permitted by law) because their fucking support takes 2 weeks to respond and they will tell you cancelling period is over.

    @sreekanth850 said: They have one of the best DDoS protection.

    ipstresser.com free package used to null route their VPS'es for hours as a "security measure" even though it was only 200Mbps which you should be able to tank even without ddos-protection as on paper their vps'es have 250Mbps connectivity. Their VPS protection is absolute garbage.
    Their "game" protection is fairly decent, but its only on certain dedis.

    @sreekanth850 said: Using their cloud instances for last 1.5 years without any support.

    Last time I successfully used their cloud services - was in 2015.
    I have attempted it recently, but due to countless billing and support related issues - I gave up, lost like $50 for a service I couldn't even use.
    So yeah, fuck them.
    I wouldn't wish dealing with them to my worst enemy.

  • So the answer is GCP or Vultr because you hate all of the others?

  • treesmokahtreesmokah Member
    edited March 2023

    @bgerard said:
    So the answer is GCP or Vultr because you hate all of the others?

    Yeah, seems like I will be going with either Vultr, GCore or if I'm really drunk - Google Cloud.
    Most likely G-Core, I had first hand experience with their support - and Its been absolutely amazing, I have no idea how much they pay them - but it must be a good sum, they are dedicated which is rare for a "big" hosting provider.

    The thread can be closed, thanks everyone.

    Thanked by 1kait
  • alt_alt_ Member
    edited March 2023

    I have experienced disk failure with Vultr a couple of times. If it was me, I would use nanode on Linode Akamai connected cloud for monitoring. It also comes with fast ticket support too.

    Thanked by 1treesmokah
Sign In or Register to comment.