(Cross-posted from Web Hosting Talk)
Over the past couple of years, I've used quite a few bottom-budget VPS providers (as in a few dollars per month for a completely unmanaged VPS).
You'd expect to get what you pay for, so many providers at that price either provide terrible service (unreliable servers, poor support when it is needed, etc) or go under (“deadpool”, in the industry speak) after a few months.
So you need to choose carefully. With any hosting, the advice is to start out monthly, and go in cautiously until you've tested the service a little. That applies even more so at this end of the market. Some providers’ best deals don't come monthly, so then it's a case of researching the risk involved at going in for a year, and making sure the annual amount is one you're happy to lose and write off if it doesn't work out.
There are a few providers that manage to make the bottom budget business model work, however. They manage their growth so they can sustain the customers they have got. They know how to run their stack, and how to boot disruptive clients that hog resources. They ensure their control panel runs well, so that support requests are for the few genuine problems rather than for routine things. They keep their network clean so their IP addresses are not blacklisted.
Obviously, at this price point, don't expect handholding customer support. If you don't know how to solve problems running a Linux server, pay another provider a bit more and they'll help when you get stuck.
Obviously, don't expect no downtime. Every provider, including high-end ones, get problems. Many excellent providers on WHT have had one or two absolute shockers in their history, but they're still highly regarded here for good reason. The issue is how often it happens, and how they handle it when they do. Do they communicate clearly? Are their team responsive? Do they honour any service level agreement?
I thought I'd help others by reviewing 8 that I've tried. All looked good enough on paper to give a go. Some worked out, some didn't. Some changed. I used these for two types of service. Some were low-resource needs, running things like DNS slaves, VPNs, or uptime monitors. Others were for storing things like offsite backups. I'm reviewing in alphabetical order, not in order of preference.
(Click on the heading of any of these providers to be taken to the provider’s homepage)
AlphaVPS
Based in Bulgaria, owned and run by Alex Barakov, with servers also available in London, USA and Germany. I've used them for both low-demand and storage.
On the whole they were very reliable. I did start to get problems with servers going offline a little too often. I can see 3 tickets with a few hours' downtime over a 6 month period. Each time, they replied after between 1 and 7 hours just to say "try now". The 7 hour ticket was opened very late evening UK time, so I wondered if their only support (even for critical problems) is based in Bulgaria. But, for the most part, they were pretty solid.
In the end I ruled them out for storage, not because of this, but because their ratio of bandwidth to disk space is only 4:1. So, if you get a VPS with 256 GB of space, you get 1 TB of transfer a month. That's fine if you're at the cold storage end of things, but for transferring daily backups offsite (then rotating them on the server) I'd have needed to order a plan with more space than I needed to get the bandwidth I required.
BuyVM
I've used BuyVM since Autumn 2011; they only started in 2010, so that makes me one of their earliest clients. I cancelled my final service with them only last month, so that's a 14 year history, but I may well use them again.
They were started by Francisco Dias (“Fran”) who is based in Canada. Originally, services were in Las Vegas, then they opened in Buffalo NY, and then in Luxembourg. They had problems with their peering in Lux, so they recently closed that colo and opened one in Switzerland. I see they've now added Miami as well.
Back in the day, they were solid as a rock. Things went wrong, sure. At points, Fran hired some quirky staff to help with billing or tech support, but even then they got the job done. You hardly needed to contact them, but if there were problems the communication was excellent, they learnt from problems, and even under pressure the tickets were responded to in a really genial and warm way. They managed this by controlling their growth. Frequently, servers are out of stock. It's always been so, even since the beginning. You have to wait for a restock, because they won't take more services than they can support.
There's lots good to say here. They've built their own VPS control panel, which is really good to use. They now offer all their services as slices (any multiple of 1 GB of RAM and 20 GB of SSD space), and slabs (multiples of 250 GB storage you can attach and detach from any slice in the VPS control panel). If you own multiple VPS in different locations, you can configure anycast. Their DDOS filtering is very good if you want to pay extra for it.
In the last couple of years, things began to change. Tickets were responded to less promptly and less courteously. I also began to notice more and more CPU steal on my munin charts, which means I'm having to wait for other clients' VPS on the same node before my tasks will run. That suggests more crowded / less well managed nodes.
The big disappointment to me was earlier this year (January 2025). Fran had begun developing a new service called NameCrane. He was clearly giving more of his time to that. Then he sold BuyVM to a company called Cloudzy. But didn't tell his clients. That's a big no-no in my book. Providers change hands, but absolutely key is how the transition is managed and how the transfer is communicated. There was zero communication. There was an announcement thread over on lowendtalk, but only certain people would notice that (I didn't). I eventually came across it. For Fran to leave a brand that's been built around his expertise and care, and not to tell the clients or to say goodbye, is really poor. I don't know anything about the new owner and operator, so no longer feel I'm in the same safe hands I've enjoyed for 14 years.
DediRock
I've only been using these guys for a couple of months, so this is an early review. I'm using them for a VPN. The actual business name is Atlas Cloud LLC, but they trade under the DediRock brand. They are owned and operated by Michael Newman Jr. They are based in Florida, with servers in Buffalo NY and Los Angeles CA.
They are currently marketing hard, with lots of special offers running, trying to build their client base and their name in the low-end VPS space. That concerns me, and time will tell if they expand faster than they can support.
So far, I've had zero problems. Servers are rock solid. TUN / TAP and all that stuff works. My heart sank when I signed up because I couldn't get VNC to work in their control panel. I feared this would be a feature I couldn't use because they hadn't set it up, but I opened a ticket to ask them to look at it. They told me they were working on it, then a couple of days later they reported it was fixed. And it was. Happy with that.
HostBrr
This is owned and run by Philip Fjaera, and I signed up to use one of their storage VPS. I'm sure they had a website of their own when I signed up (a year ago), but their main domain now just redirects to their Blesta client area. I cancelled and claimed a refund fairly quickly. I had backups to transfer from an old server I was going to discontinue using (about 200 GB worth of data). Before I started that transfer, I tested the server for its write speeds. I don't expect blazing fast on a storage system, because they will be using HDDs and because ultimately data only has to get there. But a day's backups do have to get there before the next day's run needs to start. Running a dd test to write 1 GB in 1 MB blocks, I could only get 3 MB/s write speed. I decided it was just going to be too slow.
If that's not a problem for you, though, they could be really good. I just didn't test them for long enough to confirm that. When I asked for the refund, Philip responded and refunded immediately and without any fuss, which suggests a responsive provider with integrity. At that point, I'm not going to be a long-term client, so he has no incentive to respond quickly, but he did.
InterServer
This is a longstanding provider (they've been in business since 1999), and operates out of a data centre in Secaucus, NJ. I signed up for a storage plan. Be aware they have plain storage (a DirectAdmin shared hosting plan with lots of space), and they have a storage product line under their VPS range. I wanted a full-on VPS with root control, so went for the latter.
I've been using them for 4 months now, and I've had no problem whatever. Sometimes CPU tasks on the server can be a little slow (if I need to update the kernel through DNF, this is often my slowest server to complete the task), but it gets there, and the CPU usage graph in Munin is clean. (I have – presumably shared – use of one core of Intel Xeon Platinum 8173M CPU @ 2.00GHz). You only get twice your storage space as transfer, but I did the maths and decided that their smallest plan gave me 2 TB which was enough. Their control panel feels really dated and clunky, and it's hard to find your way around, but everything you need to do is there.
What I see as a nice touch: By default, they close port 25 outbound on all servers you get from them. You need to open a ticket to request it, and give the reason. That's really sensible, as it will help to keep spammers away from their services, and so help keep their IPs clean.
RamNode
This is another provider I've used for years. They opened their doors in 2012, and my first service with them was in Feburary 2013. I've used them for both storage and low-powered needs. I stopped using my last active service with them a few months ago, but I still have the account open, with cloud credit on there, and I'll use it whenever I need to spin up a VPS to test things or to launch things.
They were created by Nick Adams, who ran it without any external / VC funding until the business was sold to InMotion in May 2021. This was a sale that was handled correctly. The sale was clearly communicated, and since Nick stepped back I've not noticed any deterioration in their service quality or the customer support. That is really rare. Sadly, more often than not, when a hosting provider is bought out from its original founder, the quality almost always deteriorates. I have no qualms at all to continue using them.
A few years ago now, they launched an OpenStack cloud service. The idea is that you sign up to that as an additional service in the normal WHMCS client area, and that creates a login to the Fleio control panel used to manage their cloud system. You then use the regular client area to raise invoices to fund your cloud account. You can then buy a wide range of VPS in the Fleio panel, corresponding to the VPS range they had available directly (regular VPS, a focus on higher storage, better premium with better RAM and processor speed, etc.). A bit similar to BuyVM's "slabs", you can also buy cloud storage (in any multiple of 1 GB). This is more flexible than BuyVM's system, but it's also more costly if you need the 256 GB minimum size that BuyVM offer, so this is better for smaller storage units. You can attach the storage to any active VPS, and it just appears as a mounted device. You can delete the VPS, but the storage remains there. I would sometimes use that for the home partition on a server; I could delete the server when I didn't need it, but at any point I could recreate the server and have my home directory back. They also let you back up any VPS to a snapshot, and you can restore that to any VPS (in any location) provided the new VPS has at least as much storage as the original.
All of this is billed by the hour, so you can leave a server running for months or years (I did that), or you can spin something up to work on for a couple of hours and just pay a few cents (I did that too). Snapshots are charged per GB per hour, and the same with storage. They honour their advertised monthly prices by capping the price at 28 days, so you still pay the same monthly rate when there are 31 days.
A couple of years after they started the cloud system, they discontinued selling their traditional product line, so now everything has to go through their cloud platform. That meant the loss of having some good annual rates available, but other than that the cloud pricing is competitive and so this is not a problem. With this cloud system, they are in the same space as companies like Digital Ocean.
RareCloud
They are owned in Cyprus, but have a UK company (owned by the Cypriot entity). They initially launched under the brand VPSforVPN, but rebranded one year ago to rarecloud.io. I signed up to use them to run a Wireguard VPN.
Initially, everything was fine. But then I hit a problem. I rebooted the VPS a few times while I was setting it up, but a kernel update a few days later made me reboot again, and it never came back up.
If I tried to use the WHMCS client area to do a manual stop and start, all was fine with the client area until I tried to access the service details page for the VPS. Thereafter, every page on the client area failed with 504 Bad Gateway. I couldn't run "Stop" or "Start". I couldn't open a ticket. I had to wait about 4 hours before I could get back into the client area. This time I opened a support ticket before I got frozen out, and trying again got me back into the 504s. It took them 6 hours to make an initial response to a "service unreachable" ticket, which is way too slow for a critical problem, and all the response said was "try again". Nothing was any different, and the only way to say I still had a problem was to hope their WHMCS email piping was working (still no client area for me).
I then got a really strange reply from them:
“There was a temporary issue with the London location. If you would like, you can try our new Tokyo location, which is fully operational.”
I explained that I ordered because I wanted London, so if the issue was only temporary, and if this only "was" an issue, I'd be happy to wait. How long did they think it would take? I also asked (again) why this was affecting my ability to use the client area. They ignored both my questions, and simply offered to recreate the VPS in London instead.
That made no sense to me. If the issue was with London, that would be no better. If the issue was not with London, my VPS should work, or they should establish corruption and apologise for it. And they still didn't acknowledge that their client area was also not working as it should.
They have a 14 day refund period, so I submitted an immediate cancellation request, and created a billing ticket to ask for the refund.
The story was not over. The refund ticket was assigned to their “senior administrator”, where it sat “on hold”. Eventually, I began joining a small chorus of others raising their refund tickets in a more public space. After 2-3 days, that got noticed, and I got my refund. Here's the fun bit: The ticket was still on hold. After a week, I got a reply.
We've reviewed your request, and your VPS has been delivered and fully active with all the resources you selected at the time of the order. At this moment, a refund is not the suitable solution, but we're here to help you get exactly what you need.
Please tell us what issue you're facing or what didn't work as expected, and we'll do everything we can to fix it for you.
In the meantime, we can offer you the following options: 1. VPS re-deployment. We can recreate your VPS in the same location or move it to another available location, keeping the same resources. 2. Refund to account balance (credit)
Rather than say that I'd already had a refund, I thought I'd see how this goes.
Hi, and thank you for this reply. Here's the thing, though. Your terms and conditions don't say that 14 day refunds apply if your team decides that to be a suitable option, or if the VPS was not delivered with the advertised specification. They simply say: “RareCloud offers a 14-day refund policy, allowing refunds within the first two weeks of service, provided conditions such as minimal usage are met.”
In response, they assigned it once again to the “senior administrator”, and I'm still waiting.
All in all, this is an avoid. I also found another big red flag after I requested my refund, but no need to share here. I think they've run promotional pricing to expand and can't now keep up with their growing number of clients. It looks to me that they don't know how to manage a VPS node. And something is very wrong with their WHMCS if an error on one server in one data centre makes the whole client area lock up for several hours. (The client area is not hosted in their London location).
Servarica
This is a Canadian owned and operated company, started by a guy called Hani (don't know his surname). It's now expanded beyond just him and they have a decent sized team.
I use them for storage, and they have some very good storage plans. You can choose between regular storage, storage where your disk quota expands a little each day so your server can grow with your needs, or storage where there is an NVMe system disk as well as a storage disk. I just went for the regular storage. The past two Black Fridays, they've run some amazing promotional plans only available at that time of year, but that come in limited quantity, so watch this space.
I've been using them for just over a year, and have had no problems at all. In that time I've only had to open one ticket. That was a slightly weird problem, but they fixed it, where the amount of RAM available to me in the XEN VPS was significantly below what the plan said. They explained that this is because the server also uses some shared memory on the node, and what I see is what's left. I'll be honest, that's beyond my understanding. I'm sure they're correct, but I could also see that I was running a kernel of my choice fully within the VM, so it looked like I had to use the memory I could see to run everything. It wasn't clear to me what memory needs were being taken off my VM by the parent node. If someone can explain that, I'd love to understand. But they increased my RAM by a couple of hundred MB to make the amount I could see correct. All solved.
I'll keep using these guys. It's quality, and it just works.
That's all folks
That's my budget VM journey. Do comment below if you've got experiences of any of these providers, or if you think I'm being unfair on any of them. I hope this has helped.
Recent comments