I preferred them to DO as they had newer (AMD CPU) hardware but then had a very bad experience with them as a customer, causing me to move my business away back to DO.
Interesting. Looks like more than half price. Are their interface / setup as easy as DO droplets? What's the catch? Are DO just expensive because they are market leaders?
They have developed drivers for Kubernetes though (CSI, CCM and Cluster Autoscaler). There are some good projects on GitHub to set it up with Terraform. I guess a real managed service will come at some point.
I recommend it as well, and one added benefit is you're in the same network as their cheap "storage box" servers (AKA SMB/CIFS/WebDAV/SSH storage server), so you can connect the two and get faster block storage compared to combining two clouds.
I have been using Backblaze/Wasabi. If you get their European servers, you can rclone mount on a Hetzner VPS and get very good speeds. If you have a fiber connection/very fast disks at home then maybe it will bottleneck. But for file storage and basic playback (for example, through SMB mount) then it is pretty good.
Tbf though, Hetzner auctions with 6TB*2 of disk at 45EUR are a good alternative. It is really a trade-off between having a server with decent power or having the flexibility of scalable S3 storage.
I wanted to build something Plex-like so sadly real block storage was a requirement, otherwise indexing and seeking could be a pain. If not I would use Wasabi or other S3-likes
Ah, I use it for Jellyfin. The only issue that I have found was storing the config files on a remote system. I have actually had that problem locally too (when using a HDD mount on a HyperV VM). But storing the config locally seemed to help.
I don't do transcodes though, and I think that is the other problem with a low-power system.
Have used them for a long time (I had their 2EUR VPS for ages, let go of it a few months ago but they honoured the price even when they were charging 4EUR for that spec). Very good customer support. I am a little worried about future prices though with electricity costs in Germany (although they have VPS in the US and Finland).
Contabo and Netcup are alternatives. Contabo has options in the UK/US/Singapore but is also German. Netcup is German, and I don't know where their servers are. OVH is another one but have had mixed experiences with their customer support.
I was affected by that fire (lost a VPS), but everything was up again a few hours later. All I had to do was to order a new VPS and restore my (very regular) backups.
The key is to be ready to recover from failure. OVH had a fire, but disks, backups, network, power, etc, can (and do) fail on premium providers. While this may say something about OVH, the effect on you as a customer is essentially the same: downtime, data loss, etc.
My experience with OVH until that point was good. 4 years without problems. I had another VPS and one dedicated server - different datacenters to avoid the "all eggs in the same basket" problem - and they were fine. The dedi, for example, I kept it from early 2017 until 6 months ago when I moved to Hetzner (better hardware, same price) and never had any issues.
I understand that not everyone can or wants to do things this way, but sometimes the savings are huge and that allows you to have more redundancy or to save money. For example, one of my websites uses a bit of bandwidth... last time I checked, it would cost around 2k/mo to run it on AWS, but only ~90 euros/mo on OVH/Hetzner. I can have another 2 replicas/servers ready to go in case of a problem and still save money.
And it's not only servers. For example, I was using S3 to store backups. Now I backup to providers like Backblaze B2, Wasabi, etc, at the same time and still save money. When I had to restore the VPS lost in the fire, I did it from my Wasabi backups because they don't charge for egress (fair use). Overall I still pay less, have more copies in different datacenters and providers, and don't have to worry about the costs of restoring backups (at least not as much).
At my company we used Contabo servers. They were Windows VMs, I can’t tell anything about Linux ones. But, the ones we had felt underpowered for what we expected. It might have been subjective, we haven’t performed any benchmarking, just a feeling. Also, their UI is straight from the 90s, in a bad way.
It doesn't take into account though:
- update guarantees
- quality of customer support (I had very bad experiences with Scaleway for example)
- score normalisation across many instances, to compensate different loads on VM hosts into account
I tried to create an account and got the following error message:
Invalid characters, allowed are: A-Z a-z 0-9 ä ö ü ß Ä Ö Ü ^ ! $ % / ( ) = ? + # - . , ; : ~ * @ [ ] { } _ ° §
What year is it, that we cannot use &?
I switched out of digital ocean because their docker apps don’t allow persistent access to the disk and don’t allow you to add any disk.
This eliminated a whole bunch of apps. I real head scratcher.
Render.com works a lot better for me. The disk is symlinked in automatically to ‘/var/data’ and I can add it in with an environmental variable set in the dashboard.
I have no idea why digital ocean cripples its app offerings like this. If anyone works there see this please add disks to docker apps.
I've been using both DigitalOcean (~5years) and Hetzner (11months) and I can definitely suggest Hetzner, very good service and good pricing - plus they offer dedicated servers at very reasonable pricing (and even vSwitch! You can create private network between them - very useful when you need a beast server and another one not too powerful).
The servers (I just tried booting the second-cheapest one) do seem much faster than DigitalOcean's and in my case they're actually cheaper, even without the coming price increase. However, one thing I noticed is that you have to pay extra for a floating IP (€3 + VAT) which is AFAIK free with DigitalOcean—or at least I didn't notice in on my bill. Something to keep in mind. I will probably switch over anyway.
Just switched everything over. I can live without the floating IP for now. I now pay slightly less than what I did at DigitalOcean while doubling the specs. Thanks for the tip!
The only problem is that now those other comments made me consider Contabo...
I'm using Oracle Cloud always free tier, and I've got 4x VMs each with 1 Arm CPU, 6gb memory and 50gb disk. Free forever. You get a small amount of object storage and some free database (nosql/sql) but I'm mainly here for the 4 VMs to run a completely free k0s Kubernetes cluster.
Oh! They have increased the prices! I went to the page hoping price reductions. Looks like inflation is changing the dynamics for the hosting business too.
The $4 droplet is a new option that is similar to what their current $5 droplet used to be some years ago. DO has never had a $2.50 droplet to my knowledge.
EDIT: I just saw your comment clarifying that you were speaking about Vultr. That wasn't clear from your phrasing, but I wanted to acknowledge it here.
Honestly I think enticing shareholders is one of their primary motivators here, and the inflation is more of the excuse. Nobody else I work with is raising their costs in this space right now (we'll see how long that lasts though), and DO isn't exactly the bottom of the list on pricing.
The increase doesn't look too large, but if you need a similar alternative I like Vultr a lot, the Neocities CDN is hosted there and it's been rock solid performance for years, and it's quick to get access to competent tech support. I wish I had more infra hosted there actually.
Hosting is naturally deflationary, since the cost of the underlying hardware becomes cheaper per unit over time. Take storage, which has gone from roughly 10 to 2 cents per GB over a decade [0]. I believe there's a quote somewhere about AWS initially selling EC2 unprofitably, with the expectation that their costs would go down.
Not sure the underlying economics here for DO or their motivation, but I don't think DO can expect to their customers to react to this the same way customers are reacting to inflated costs in cars, groceries and the like. If other providers follow suit, maybe that's a sign of slowing efficiency in the hardware, end of Moore's law, etc.
Changing the dynamics how when it is already overpriced. This is like all the businesses during COVID saying... It's COVIDs fault our customer service is terrible.
I think it's overcharging customers to see how much they can get away with, like netflix. Jump on the "iTs inFlAtIoN" bandwagon and see how much profit you can squeeze out and see how well you can handle customer attrition while doing it. If it goes bad just say "oh looks like inflation is lightening up a bit now". I call it the Netflix Gambit (circa 2022)
I mean even with the price increases my little server that could that serves some basic self-hosted apps up for me still costs me less that a meal at a non-fast food restaurant. I'm sure I'm getting more utility out of that than 1 meal
How many places show a difference in their prices? How often to restaurants, stores, and other as a service companies doing that? I almost never see it. I wouldn't say they are doing something deliberate (in a negative sense). They are doing something pretty typical.
Restaurants and stores are single transactions, not ongoing subscription services.
DigitalOcean makes you dig for the price increase, which is pretty scummy. They tried to soften the blow with a new droplet which looks even worse. At the end of the day, 20% is a huge change.
They have a table of everything listed along with a new price for it. All they would need to add is a column showing the existing price. There is a 100% chance they have a spreadsheet internally passed around with that existing price column.
Do you really believe it was a not a deliberate choice to not show the side by side comparison?
> That’s why today we are announcing that for the first time we’re changing our prices to best serve you at every stage of your growth journey.
To best serve you at every stage of your growth journey it would seem. Or if you'd like a translation from stilted corporate speak into plain language: we went public and the line's gotta go up and engineers and hardware don't come cheap.
It more and more feels like going public is the big mistake in a lot of tech companies, because everything gets real squirrely after that for customers and employees.
Any more, whenever a company I use goes public or is acquired, I start looking for the egress. I'm happy that the founders of whatever it is got a nice payday, but I don't expect their creation to continue being worth using after that.
the thing is once you start private and giving options as comp, your only option is to go public eventually, or screw everyone that's every gotten options.
Profitable private companies that never want to go public can solve this by doing profit sharing with employees, but in tech there's often no profits for a very long time.
I agree. I understand pricing increases (even if I don't look forward to them) and I'm an adult, just tell me the difference in prices instead of deliberately obscuring that information. Not showing the prices smells of not standing beside your product/new pricing. I think DO is still worth it at the new prices, that said I'm having a hard time remembering when a cloud provider raised prices like this, normally the trend is down/same no?
From a very quick look I think the cheaper basic droplets have gone up 20%, but larger ones (such as the "general purpose droplets") have only gone up as little as 5%.
Edit:
Yes, it's only "Basic" droplets that have gone up 20%, everything else is only 5%.
I actually got slightly burned when they did that update. I was doing deployments by "slug" (their term) and the old slug went up to $10/month when they created the new 1GB $5 droplet option. Now my system double checks the price before deploying a droplet.
The battle between cheaper hardware and inflation was something I found interesting and wondered which would affect prices first. I personally won't change anything, I'll pay the extra $1 a month for my cheap droplet to not have to move anything. DO's big selling point to me (as a hobbyist) was the simplicity. Now that I work with AWS every day, I feel very comfortable getting some basic infra set up, but I still really like DO's smooth setup.
It's a shame they had to do it. I wonder how difficult it is to compete with all the big providers? They still fit my use case just fine (still a big cheaper than the AWS equivalent too I think, but maybe not) and I'm fortunate to not be in a situation where the extra dollars are a deal breaker.
Interesting. While it doesn’t seem that unreasonable, this change would really add up if you already had a lot of droplets, and I kind of did. I suppose it’s good luck that I’ve been migrating workloads to Fly lately. Maybe I should finish the job.
To be fair, once you pass the Fly free tier, I can only guess it is more expensive, but in my case it’s made up for by being able to do much smaller VMs. Digital Ocean’s a bit more limited on how you can cut up memory and disk space (and the dynamics of k3s or kubernetes at this scale is awkward.)
I operate a service called RunBuildRun that sells managed GitLab runners that run on DigitalOcean droplets. I have lots of droplets. I may need to bump up my price a bit :)
Sorry for the tangent, but can you provide more info on your service? Without signing up from an account on your site, I'm unsure on how the interface here compares to GitLab runner registration (which I always felt was a bit convoluted due to how GitLab structures their user and organization accounts) and how the hardware compares to what GitLab offers.
RunBuildRun makes creating a runner a one click operation. We give you a registration token for your runner, which you then register in the GitLab interface like usual.
The primary value over GitLab's runners is that RunBuildRun doesn't have any CI/CD minute quota. With GitLab, you get a finite quota (e.g. 400 minutes on the free tier) and then you have to upgrade to a per seat premium plan or buy additional CI/CD minutes.
Also, RunBuildRun gives you a dedicated runner so you don't have to wait on shared runners and you don't have to setup/manage a runner on your own server.
The US government admits to 8% inflation in the last 12 months. This price increase seems completely fair given the number of years it's been unchanged.
Interesting (or maybe not?) that the App Platform and Spaces pricing hasn't changed. I've been migrating as many projects as possible over to App Platform + Managed DB + Spaces recently. Although it's been frustrating due to their documentation being lacking, I've found it great for simple projects and static websites once you get your head around it. It's certainly more expensive than running plain Droplets, but the convenience is worth it. Their costs must be better covered on the App Platform over Droplets.
I was a big fan of Nanobox (the company DigitalOcean acquired and then turned into App Platform). And Pagoda Box before that. I've never found any PaaS that was as easy to use as those — especially because it came with a local environment for development that would match the production environment.
I'm very interested to try App Platform for my next side project. In fact, I still have a server running on DigitalOcean that was set up through Nanobox a few years ago. I need to move that over to App Platform at some point.
It might be different for you having come from the predecessor, but I found it quite hard to get started with the App Platform. There is a conflict between DO's attempt to have a web interface that allows you to create and manage apps, and the underlying YAML/Buildpacks that actually model your apps. While the web interface is OK for seeing your apps at-a-glance, it's definitely not OK for creating and configuring new apps. For that you need to use the YAML config files and probably doctl (their CLI app). It took me a long time to understand this. Their docs don't do a good job at explaining the App Platform at a high-level. That said, I've still found it very useful for "simple" webapps.
It’s always nice to hear that someone liked Pagoda Box. I worked there way back in 2011 before the Nanobox change, and it was really fun working with containers, dynamic load balancers, and the `git push` workflow before those were widely known.
Nothing was as magical as having a site up a running a few seconds after doing `git push pagoda`.
App Platform is a great concept, but we hit a dealbreaking road block when trying to migrate some Python apps with job queues. Their runtime (gVisor) doesn't support semaphore locks, which is used by Pythons multiprocessing and in turn used by most job runners (we discovered it with django-q, but I think most, if not all of them including Celery, rely on this, see link below).
The build times for Dockerfiles are also atricious… our build failed after 40 minutes by running out of memory and the multi-stage Dockerfile really wasn't anything special. We would have just used the images hosted on Github Container Registry, but App Platform only supports a limited range of Docker registries too. Note: the images build in 3 minutes on Github Actions.
As far as I can see it is also not possible to add any block storage too. While I mostly work on projects that use object storage anyway, SOME things just need persistent block storage. Which is annoying, since DigitalOcean HAS block storage… just not for App Platform.
I really wanted to use it, but man they make it hard.
Seems like the $4 droplet is a smoke screen hiding price increases. I'm a long time DO customer. I wish they would have just said "prices are going up" and listed a comparison of the old price with the new increased price. In their new price list, they do not list the old price... so you cant tell if they are going up and down. Because the article starts off with the $4 droplet announcement, you get a feeling that they have made changes to their infrastructure and prices will be going down to pass along those savings.
Quite frankly, I feel this only happens when we're in a time of wealth and growth. I suspect many more companies will start acting this way. It's simply a logical conclusion based on the current market sentiment and the view that I outlined above.
I was going to say that the email subject said prices "are going up", but it turns out the subject line was "DigitalOcean prices are changing July 1st. Here’s what you need to know." and I just assumed changing went up so strongly I remembered it wrong.
Seems like a really reasonable price increase (though the timeline is short), but they ought to have been a little clearer about exactly what the before-and-after picture was.
Ive quit them long time ago when they did not adjust price to compete. They also had for a few years, inferior performance per dollar as others in same range.
20% bump for the most budget-sensitive option is a crazy increase. Sure it's fine if you only have one, but I'm sure there are many folk who have hundreds
$100/hour+ human costs to migrate though right. Imagine downsizing 100 web apps, notifying clients of what's going on, and dealing with random breakages. That's not an hour's work
You can get a GCP micro VPS for free, “forever”. I have a free VPS always running, and run beefier VPSs as-needed for my personal projects. For my needs, GCP and Colab (sometimes Pro, sometimes free) are convenient and time-savers.
You can also get a free Oracle Cloud VPS, and Alibaba provides good cloud value also. Many years ago, I enrolled in Microsoft’s BizSpark program, and if they still do that, check them out. So many good options…
Oracle gives you 10TB of monthly bandwith + a 4 core/24gb ram ARM VM + 1 or 2 (can't remember) 1Core/1GB x86 AMD Epyc core... for free. If you are just looking to get a cloud VM for personal projects, it's by far your best option.
It's a bit weird to say considering it's oracle, but there's no catch and no hidden costs. As long as you don't upgrade to a paid account, there's no way to even mess up and get charged
Is that really the case? I just ran the CLI version of speedtest.net and I got way more than the 4Gbps that oracle advertises. Though I only ran the test on ARM, and I know AMD VMs get much less bandwidth. I also wouldn't be surprised if my results aren't representative of real world sustained usage.
I've priced up DO as an alternative to AWS on several projects and it has never been cheaper once AWS discounts (reservations, spot) are applied correctly. You need to consume a relatively large amount of bandwidth before DO becomes cheaper, and even then a third party CDN is likely to be a better solution
The $5 droplet includes the equivalent of $90 worth of AWS egress, although it is accounted somewhat differently.
An equivalent low usage machine on AWS in us-east (t3a.micro - 1GB / shared CPU) on spot instead costs around $2.16, almost 57% less.
So this is where bandwidth becomes interesting, especially in combination with an external CDN.
Say if you're hosting 100 low volume client sites where each client gets its own instance (sensible security choice), DO you'd be paying $284/mo. more than equivalent AWS config, or you could treat the $284 as a bandwidth allowance, leaving 3.155 TB egress in aggregate across all clients before DO/AWS are breakeven again.
If some/many of those client sites consume a lot of bandwidth, the equation quickly changes, but you need to start looking at edge hit rates of an individual application before attempting to make a meaningful comparison.
The last project I looked at this for included several high bandwidth sites with good edge hit rates on an external CDNs, and AWS still came in significantly cheaper
This is not true at all. Electricity prices are crazy and hardware is a lot more expensive for us (not affiliated with DO, I run a PaaS service) than it was 2 years ago.
I find it particularly amusing that they're reintroducing the 512MB of memory tier, but at $4 instead of the $2.50 it was before the removal. But I suppose that's roughly inline with the basic 1GB/RAM droplet going to $6. I'm most interested to see how the managed Kubernetes offering changes shake out -- hoping I don't sense a migration in the near future.
I think this price increase is honestly pretty fair, but I really wish they'd show side by side prices, and offer me a tool to see what my total increase will be.
Because Digital Ocean is a company with employees and investors, not just an automated conglomeration of servers. A 20% increase in prices after 12 years of basically stable prices seems fair.
I just checked one of the node types and EC2 is about 20% more expensive, even after the price hikes. AWS may have more baked-in margin than DO.
Also, I didn’t include Transfer costs in the cross-check, so AWS may be much more expensive.
Depending on your needs, AWS is hands down one of the most expensive ways to put workloads into the cloud.
At the same time it’s a bit difficult to compare them apples-to-apples. In my experience EC2 instances aren’t designed for reliability as much as they are designed to meet exactly what AWS can put on the invoice as being the value exchanged for money. I’m not bashing their design goals, they’re just different from “traditionally VPS-first” firms.
If you’re running not-huge On Demand instances you can definitely get much better performance for your dollar on something like Linode or DigitalOcean, OVH, Hetzner, etc. And then you can still some of the AWS services where they have much less comprehensive competition. IMHO, you can beat EC2 all day but feature-wise it’s extremely difficult to beat S3.
Seems like AWS has really only gone down in price on the pre-provisioned EC2 instances, not the on-demand. On-demand is more inline with what Digital Ocean provides.
I disagree, CPU performance per $ is getting cheaper. Increasing pricing on the $5 droplets by 20% is significant and their deceptive email is also an issue. I'm going to cancel my droplets.
One of the upsides of only being 50% cloud is that I get some perspective.
The cheaper, better, faster assumption is being challenged in most segments. Shipping costs have been crazy and eating margins for just-in-time delivery for some time. Containers are 5x-10x from China vs 2018, a tractor trailer costs $1400 to fill with diesel today in my region.
Also, commodities like storage aren’t getting cheaper.
I don't just pay DO for servers, I'm paying them for a cohesive integrated UX, decent support (for the price), and nice hosted database and load balancer options. That's always been what separates DO from the rest of the market. Servers are cheaper elsewhere, that extra stuff is all pretty competitively priced.
All of those things are subject to the prices of people, which are nuts right now.
I also have deployments at hetzner and OVH and the management angle is day and night (although Hetzner is almost there now). Multiple different logins and dashboards by product, country, etc.
I'm sure some of it is just driven by being public as well. I'll keep paying them if they can keep those other factors great, but little things like wasting my time on the pricing table above are damaging the one thing that I really like about DO.
I’m not saying you don’t - but my experience has been different. In my experience Linode has customer service that is in a different league than DigitalOcean has ever provided. I’ve never gotten what I would call great, or perhaps memorably good, customer service from DigitalOcean. I haven’t gotten awful customer service either. It’s the kind I would expect from any turn key solution that doesn’t employ support agents who are already well versed in what they’re working on.
Yes, the price of computation is going down, but the prices for real estate and electricity are going up, often way up. Computers use lots of electricity and sometimes the price of electricity over the lifetime is more than the cost of the machine. Gasoline prices don't track electricity prices in the short term, but eventually the prices of energy catch up with each other. They can't escape inflation.
It is not like hardware costs are the only costs. My electricity costs are going to be nearly doubling shortly. They could be seeing the same thing. I wouldn't be surprised if employee retention/turnover is another issue they are facing.
Computation needs electricity, computers need maintenance, and the maintenance is done by people with salaries. People have to drive to work which uses gas. Parts of computers have to be shipped, which also needs gas.
Gas and electricity are going up.
You have no clue what you're talking about. Computation prices are not going down.
Customers use DigitalOcean for use cases ranging from hosting mission-critical applications to testing new ideas. To empower even more developers across the globe, we are introducing a new Droplet priced at $4/month with 1 vCPU, 512MB memory, 500GB bandwidth, and a 10GB SSD.
I'm kind of annoyed that vCPU is now a measurement of ... something.
Whereas memory, bandwidth, and storage are "roughly" comparable (there are performance differences between SSDs and RAM for sure, they're not as blatant) the difference between "a single core" on a CPU can be completely insane.
We need some form of "generic CPU power" that can more accurately compare these things.
Just closed a bunch of droplets. This is a deceptive price increase hidden behind a bunch of annoying PR. I guess it's understandable with everything going on but the market is beginning to get very hostile towards smaller startups with a lot of smaller services.
I don't know if I would call them the 'best', but Kamatera offers pretty good bandwidth, 5TB/mo or 50MB/s unbounded at $4 (for non-dedicated allocation, $9/mo for dedicated). I don't know if I'd call them the best because I don't really need much more than something that's reliable and has decent speed, which it satisfies pretty well.
I’ve been paying for a cheap dedicated server for ages for some things but never bothered to migrate my DO workloads. With the price increase this means my DO workloads cost nearly as much as the dedicated server, and I know they can be containerized and shifted easily.
To give some hobbyist level price increase numbers:
I'm currently running two basic droplets at 8gb RAM and 4vCPU, which is basically just enough resources for a proper dev kubernetes cluster with Promethueus, Loki, 5-6 apps, etc (4gb 4vCPU is not enough, I tried). I also run a third 1gb1vcpu management droplet outside the cluster as well as a load balancer in front of my API gateway and a spaces instance to dump backups and store stuff like terraform state. My previous bill was $102 and it looks like it will now under this new pricing with all of the above be about $125 which is about 20% increase.
Are there similar managed kubernetes offerings out there that are better on price? My understanding is that I could probably half this or even do better if I run my own control plane on some other provider but if I want managed k8's is this still the best option price wise? I guess if something like Hetzner or OVH is less than half the price it may even be worth the headache to run my own control plane (which has it's own advantages too - currently I'm stuck with Cilium on DOKS and it would be nice to use another network provider for some of my use cases).
Yep - it's also an exercise to keep my cloud engineering skills up to date - I can take the approach I use on the hobby projects and apply it to production level stuff in my job.
If it's an exercise to keep your cloud engineering skills up to date, perhaps you'd like to manage your own Kubernetes? Then I'm sure it can be much cheaper.
Spending money on hobbies is part of the benefits of working no? If you can afford the essentials I would hope people are spending money on things they enjoy.
Very very low - I used only 30 gb of the massive amount DO gives you last month. Basically currently just have some stuff periodically sending data in and most of the outbound is just small REST responses.
You might want to look into OVH's k8s offering [1]. You only pay the worker nodes and the general purpose instances cost 26,18 € per month offering 2vcpu and 7 gb RAM [2].
This is Raspberry Pi 4 territory. Your monthly cost is the price of hardware, per month. If you are on a home 1G connection, your home bandwidth is as good or better than what you are getting on DO. At the end of a year, you would have a 12 node cluster.
PS Shame on you Radxa for naming your module CM3, shame. There will be a flood of RK3588 based devices on the market in the next 6-12 months. A low end PC would replace 10s of these devices.
My point was that a 4 vcpu instance is overpriced and also a very small amount of compute. Most cloud costs are 1/3 to 1 of price of hardware, per month.
I am not advocating that people run k3s on raspberry class hardware, just that GP is overpaying for what they are getting. An Ebay server machine colo'd would be several orders of magnitude more capable at the same cost.
My 1G connection sees 250M up and down continuously.
Is it important that they're losing the branding of the $5 droplet? It seems that's significantly more of an impressionable number than 4 or 6, in my understanding of the factors that made them popular. To me they've simply lost the #1 thing that people remember them by and don't stand out with that simple change. That's the only change that worries me.
I thought Linode's "Nanode" ($5 VM) was several years after DO. I remember wishing Linode had a lower tier option and contemplating moving to DO back when Linode started at $20 for their lowest spec.
I went back to check and indeed my annual billing for Linode's lowest tier was was $216 annually (12 * $20 * 0.90 for the 10% annual payment discount.)
Linode then dropped the price in 2014 to $10/month [1], and eventually in 2017 introduced the $5 option [2]. DO was definitely around well before then, but I do recall DO becoming more popular around that time. Because for $5 it was awesome to spin up a VM to test a new app.
That's what I call a misleading title, my droplet is going up 20% with this news.
Sad, my hardware hasn't changed in 4 years, why is it suddenly more expensive to host? Hardware costs are always dropping, there's no reason for this other than greed.
I guess this is another "inflation" thing where companies demand higher profits without changes in services.
> Hardware costs are always dropping, there's no reason for this other than greed.
They're not though, there's a load of world happenings currently resulting in chipset shortages and increased energy pricing etc in many parts of the world.
Maybe, but that doesn't impact existing hardware. So unless this upgrade includes things like newer CPUs or more ram/disk for a droplet that doesn't apply, and as far as I can tell, the hardware I started on 4 years ago is the same. At least they've not advertised any new CPUs or increased disk or ram.
340 comments
[ 3.4 ms ] story [ 272 ms ] threadNo managed databases or kubernetes clusters though. Hoping that’s in the works as it will really level up their offering if they do.
Hetzner Cloud looks modern and it's easy to use. They don't have the same services as DO though.
If you go the dedicated server way, you'll use the old control panel. I don't find it hard to use, but it's not as good looking.
Tbf though, Hetzner auctions with 6TB*2 of disk at 45EUR are a good alternative. It is really a trade-off between having a server with decent power or having the flexibility of scalable S3 storage.
I don't do transcodes though, and I think that is the other problem with a low-power system.
I use it for all my projects for 3 years, and very satisfied
Contabo and Netcup are alternatives. Contabo has options in the UK/US/Singapore but is also German. Netcup is German, and I don't know where their servers are. OVH is another one but have had mixed experiences with their customer support.
DO prices are ridiculous (imo).
I get maintenance emails quite often, but I think it's from an abundance of caution. I can't remember it actually ever going down.
https://www.datacenterdynamics.com/en/opinions/ovhclouds-dat...
The key is to be ready to recover from failure. OVH had a fire, but disks, backups, network, power, etc, can (and do) fail on premium providers. While this may say something about OVH, the effect on you as a customer is essentially the same: downtime, data loss, etc.
My experience with OVH until that point was good. 4 years without problems. I had another VPS and one dedicated server - different datacenters to avoid the "all eggs in the same basket" problem - and they were fine. The dedi, for example, I kept it from early 2017 until 6 months ago when I moved to Hetzner (better hardware, same price) and never had any issues.
I understand that not everyone can or wants to do things this way, but sometimes the savings are huge and that allows you to have more redundancy or to save money. For example, one of my websites uses a bit of bandwidth... last time I checked, it would cost around 2k/mo to run it on AWS, but only ~90 euros/mo on OVH/Hetzner. I can have another 2 replicas/servers ready to go in case of a problem and still save money.
And it's not only servers. For example, I was using S3 to store backups. Now I backup to providers like Backblaze B2, Wasabi, etc, at the same time and still save money. When I had to restore the VPS lost in the fire, I did it from my Wasabi backups because they don't charge for egress (fair use). Overall I still pay less, have more copies in different datacenters and providers, and don't have to worry about the costs of restoring backups (at least not as much).
I'm a bit sad they don't have anything closer to the UK: I'll miss the 4ms ping to my DO VPS.
EDIT: holy shit, I looked into Contabo mentioned below and it's even cheaper! €5 for 4 cores and 8 GB RAM, what's the catch?
It doesn't take into account though: - update guarantees - quality of customer support (I had very bad experiences with Scaleway for example) - score normalisation across many instances, to compensate different loads on VM hosts into account
This eliminated a whole bunch of apps. I real head scratcher.
Render.com works a lot better for me. The disk is symlinked in automatically to ‘/var/data’ and I can add it in with an environmental variable set in the dashboard.
I have no idea why digital ocean cripples its app offerings like this. If anyone works there see this please add disks to docker apps.
"You get you pay for", even if it is half true accounting for DO's shareholder's pressure, I think Hetzner is impossibly cheap.
The only problem is that now those other comments made me consider Contabo...
Terrible support. https://community.oracle.com/tech/apps-infra/discussion/4496...
Usually - well before inflation - this tier was $2.50.
EDIT: I just saw your comment clarifying that you were speaking about Vultr. That wasn't clear from your phrasing, but I wanted to acknowledge it here.
The increase doesn't look too large, but if you need a similar alternative I like Vultr a lot, the Neocities CDN is hosted there and it's been rock solid performance for years, and it's quick to get access to competent tech support. I wish I had more infra hosted there actually.
Not sure the underlying economics here for DO or their motivation, but I don't think DO can expect to their customers to react to this the same way customers are reacting to inflated costs in cars, groceries and the like. If other providers follow suit, maybe that's a sign of slowing efficiency in the hardware, end of Moore's law, etc.
[0]: https://www.backblaze.com/blog/wp-content/uploads/2017/07/ch...
I love DO, but with the pretty basic requirements I have (just a blog and some small apps) I need to start looking at cheaper options.
DigitalOcean makes you dig for the price increase, which is pretty scummy. They tried to soften the blow with a new droplet which looks even worse. At the end of the day, 20% is a huge change.
Do you really believe it was a not a deliberate choice to not show the side by side comparison?
To best serve you at every stage of your growth journey it would seem. Or if you'd like a translation from stilted corporate speak into plain language: we went public and the line's gotta go up and engineers and hardware don't come cheap.
Profitable private companies that never want to go public can solve this by doing profit sharing with employees, but in tech there's often no profits for a very long time.
Edit: Yes, it's only "Basic" droplets that have gone up 20%, everything else is only 5%.
I haven't checked their pricing in a while but I think the $5/mo droplet used to be 512m but is now 1G ram, right?
It sounds like there's also a $4/mo smaller droplet with 512m ram.
It's a shame they had to do it. I wonder how difficult it is to compete with all the big providers? They still fit my use case just fine (still a big cheaper than the AWS equivalent too I think, but maybe not) and I'm fortunate to not be in a situation where the extra dollars are a deal breaker.
- Hitting the machine with the hammer: $5
- Knowing where to hit the machine to get it working again: $995
To be fair, once you pass the Fly free tier, I can only guess it is more expensive, but in my case it’s made up for by being able to do much smaller VMs. Digital Ocean’s a bit more limited on how you can cut up memory and disk space (and the dynamics of k3s or kubernetes at this scale is awkward.)
The primary value over GitLab's runners is that RunBuildRun doesn't have any CI/CD minute quota. With GitLab, you get a finite quota (e.g. 400 minutes on the free tier) and then you have to upgrade to a per seat premium plan or buy additional CI/CD minutes.
Also, RunBuildRun gives you a dedicated runner so you don't have to wait on shared runners and you don't have to setup/manage a runner on your own server.
I'm very interested to try App Platform for my next side project. In fact, I still have a server running on DigitalOcean that was set up through Nanobox a few years ago. I need to move that over to App Platform at some point.
Nothing was as magical as having a site up a running a few seconds after doing `git push pagoda`.
Of course I am biased though :)
The build times for Dockerfiles are also atricious… our build failed after 40 minutes by running out of memory and the multi-stage Dockerfile really wasn't anything special. We would have just used the images hosted on Github Container Registry, but App Platform only supports a limited range of Docker registries too. Note: the images build in 3 minutes on Github Actions.
As far as I can see it is also not possible to add any block storage too. While I mostly work on projects that use object storage anyway, SOME things just need persistent block storage. Which is annoying, since DigitalOcean HAS block storage… just not for App Platform.
I really wanted to use it, but man they make it hard.
https://github.com/Koed00/django-q/issues/522#issuecomment-1...
PR folks who work in tech companies need to realize that their customers are different from, say, a toothpaste company's customers.
Just be fucking transparent.
Seems like a really reasonable price increase (though the timeline is short), but they ought to have been a little clearer about exactly what the before-and-after picture was.
All that behaviour let linode and vultr flourish
You can also get a free Oracle Cloud VPS, and Alibaba provides good cloud value also. Many years ago, I enrolled in Microsoft’s BizSpark program, and if they still do that, check them out. So many good options…
It's a bit weird to say considering it's oracle, but there's no catch and no hidden costs. As long as you don't upgrade to a paid account, there's no way to even mess up and get charged
If it switches at $10 worth of bandwidth, or even at $20, is that REALLY that much?
An equivalent low usage machine on AWS in us-east (t3a.micro - 1GB / shared CPU) on spot instead costs around $2.16, almost 57% less.
So this is where bandwidth becomes interesting, especially in combination with an external CDN.
Say if you're hosting 100 low volume client sites where each client gets its own instance (sensible security choice), DO you'd be paying $284/mo. more than equivalent AWS config, or you could treat the $284 as a bandwidth allowance, leaving 3.155 TB egress in aggregate across all clients before DO/AWS are breakeven again.
If some/many of those client sites consume a lot of bandwidth, the equation quickly changes, but you need to start looking at edge hit rates of an individual application before attempting to make a meaningful comparison.
The last project I looked at this for included several high bandwidth sites with good edge hit rates on an external CDNs, and AWS still came in significantly cheaper
Fwiw, that 2.5 was after the price drop. 1vcpu + 512mb + 20gb used to be $5
For perspective, Hetzner Cloud's CX11 nodes pack 2GB RAM and go for €4.15/month.
https://www.hetzner.com/cloud
If anything I would expect it to go down as computation and memory get cheaper.
At the same time it’s a bit difficult to compare them apples-to-apples. In my experience EC2 instances aren’t designed for reliability as much as they are designed to meet exactly what AWS can put on the invoice as being the value exchanged for money. I’m not bashing their design goals, they’re just different from “traditionally VPS-first” firms.
If you’re running not-huge On Demand instances you can definitely get much better performance for your dollar on something like Linode or DigitalOcean, OVH, Hetzner, etc. And then you can still some of the AWS services where they have much less comprehensive competition. IMHO, you can beat EC2 all day but feature-wise it’s extremely difficult to beat S3.
Possibly also gearing up for a sale, and keen to test the elasticity of the market.
The cheaper, better, faster assumption is being challenged in most segments. Shipping costs have been crazy and eating margins for just-in-time delivery for some time. Containers are 5x-10x from China vs 2018, a tractor trailer costs $1400 to fill with diesel today in my region.
Also, commodities like storage aren’t getting cheaper.
All of those things are subject to the prices of people, which are nuts right now.
I also have deployments at hetzner and OVH and the management angle is day and night (although Hetzner is almost there now). Multiple different logins and dashboards by product, country, etc.
I'm sure some of it is just driven by being public as well. I'll keep paying them if they can keep those other factors great, but little things like wasting my time on the pricing table above are damaging the one thing that I really like about DO.
How much the consumption reduction counteracts the rise in electricity prices, I don't really know.
Saying it's just computation and memory is ignoring 99% of what it takes to run that business.
Gas and electricity are going up.
You have no clue what you're talking about. Computation prices are not going down.
> Introducing a new $4 Droplet
Customers use DigitalOcean for use cases ranging from hosting mission-critical applications to testing new ideas. To empower even more developers across the globe, we are introducing a new Droplet priced at $4/month with 1 vCPU, 512MB memory, 500GB bandwidth, and a 10GB SSD.
Whereas memory, bandwidth, and storage are "roughly" comparable (there are performance differences between SSDs and RAM for sure, they're not as blatant) the difference between "a single core" on a CPU can be completely insane.
We need some form of "generic CPU power" that can more accurately compare these things.
https://www.lastweekinaws.com/blog/what-is-an-ecu-an-ec2-com...
Any suggestions on the best VPS providers for bandwidth-intensive applications?
Guess I have a reason to migrate now.
https://www.hetzner.com/cloud (scroll down to unbeatable prices).
and it's been that way for a long time.
There's also: https://www.ovhcloud.com/en/vps/
... why is this news ?
I'm currently running two basic droplets at 8gb RAM and 4vCPU, which is basically just enough resources for a proper dev kubernetes cluster with Promethueus, Loki, 5-6 apps, etc (4gb 4vCPU is not enough, I tried). I also run a third 1gb1vcpu management droplet outside the cluster as well as a load balancer in front of my API gateway and a spaces instance to dump backups and store stuff like terraform state. My previous bill was $102 and it looks like it will now under this new pricing with all of the above be about $125 which is about 20% increase.
Are there similar managed kubernetes offerings out there that are better on price? My understanding is that I could probably half this or even do better if I run my own control plane on some other provider but if I want managed k8's is this still the best option price wise? I guess if something like Hetzner or OVH is less than half the price it may even be worth the headache to run my own control plane (which has it's own advantages too - currently I'm stuck with Cilium on DOKS and it would be nice to use another network provider for some of my use cases).
[1] https://www.ovhcloud.com/en/public-cloud/prices/#568
[2] https://www.ovhcloud.com/en/public-cloud/prices/#419
This is Raspberry Pi 4 territory. Your monthly cost is the price of hardware, per month. If you are on a home 1G connection, your home bandwidth is as good or better than what you are getting on DO. At the end of a year, you would have a 12 node cluster.
https://www.jeffgeerling.com/blog/2021/pine64-and-radxas-new...
https://www.hardkernel.com/shop/odroid-m1-with-8gbyte-ram/
PS Shame on you Radxa for naming your module CM3, shame. There will be a flood of RK3588 based devices on the market in the next 6-12 months. A low end PC would replace 10s of these devices.
My point was that a 4 vcpu instance is overpriced and also a very small amount of compute. Most cloud costs are 1/3 to 1 of price of hardware, per month.
And home connections are all asymmetric, you may get a gig down but I've not seen anything outside business class with over 50 up.
My 1G connection sees 250M up and down continuously.
[1] https://www.mythic-beasts.com
Linode then dropped the price in 2014 to $10/month [1], and eventually in 2017 introduced the $5 option [2]. DO was definitely around well before then, but I do recall DO becoming more popular around that time. Because for $5 it was awesome to spin up a VM to test a new app.
[1] https://www.linode.com/blog/linode/11th-linode-birthday-10-l...
[2] https://www.linode.com/blog/linode/high-memory-instances-and...
Sad, my hardware hasn't changed in 4 years, why is it suddenly more expensive to host? Hardware costs are always dropping, there's no reason for this other than greed.
I guess this is another "inflation" thing where companies demand higher profits without changes in services.
They're not though, there's a load of world happenings currently resulting in chipset shortages and increased energy pricing etc in many parts of the world.