Digital Ocean killed my droplets and gives a vague reply
Any idea what could be going on here?
I had two droplets on DO, a server for game halfchess.com and a VPN that I use at work.
I get this message from DigitalOcean Abuse. Could I have been targeted by an individual or is this something to do with a software loophole they discovered?
----
Hello there,
Thank you for your response.
We understand your concern and have reviewed this ticket extensively but we are unable to accommodate this unlock request. In order to maintain the integrity of our security systems, we cannot disclose which factors were identified that led to the decision on this account. We hope you would understand the limitations on our end. I actually moved over to Linode because of similar anomalies. I had a container I was running my personal site from. Paid my bill on time every month. Got distracted with work and hadn't updated it in a while. One day I go to check on it and the site is down. Head to DO to restart it... and it is just gone. Like it never existed. Didn't even bother trying to communicate with them I was so angry. Been really happy with Linode. I'm pretty happy with Linode after migrating from DO as well. Their billing seems a bit more on top of things, noticing situations where they owed me a small credit that I would have likely missed. I've been a customer since before the Akamai days and feel that their service has only improved when I expected disruption (as one does when companies merge). I mean, yeah, if you don't pay your bill you should fully expect them to stop providing the service which was being paid for (which includes the storage of the associated data). He said he paid his bill on time every month, he just didn't update the site in a while! Generally it’s not unreasonable to expect a business to care about customer retention… and have them email you before this happens Please read posts before replying to them. I would say it depends on how long the bill is not payed. I wouldn't expect them to remove everything when a payment is late for a month, for example. But if not payed for more than 6 months, then it is understandable. This is not unique, unfortunately. I have been battling with them for the last two months now and I am in the process of moving everything away from DigitalOcean. Something has changed over there and they have zero tolerance concerning everything. My account was overdue by $10 (after being paid on time for close to 5 years), I asked for a deadline extension of 1 week and it was declined. They have already turned off all of my droplets and are now threatening to delete everything with no recourse within the next three days. Again, not unwilling to pay, I was merely asking for an extension by 7 days for $10 and they declined my request. I don't see anything wrong with what they are doing in your case, no money no service. What are popular alternatives these days? Vultr, you can get your first month free through various referral / promotion links, where your first month’s spend (up to $50-200 depending on the link) is gratis (after you’ve added a legit payment method) Get yourself an Oracle cloud freetier. It's free and you can get 2x 1Ram1CPU x86 VM's and any combination of 24Ram4CPU arm64 VM's. There is also a bunch of other resources that you get access to. Be aware of VM reclamation, but it is easy to work around it. More here: https://docs.oracle.com/en-us/iaas/Content/FreeTier/freetier... From my experience, it's really hard to create an account and upgrade your account to PAYG. People saying you need that if you want to get the free tier VM because the free tier VM keep out of stock. I need to email the customer support to make an account and even then to change my account to PAYG account, I need to register the credit card again, they keep rejecting my credit card even though that is the same credit card that I used for my account and the test payment go through. After that I just give up and go with Hetzner (same you need to email CS) but I can start paying them immediately after the account created. I had a different experience. I upgraded a free account last year without needing email anyone. And only did that to use some non-free feature. I had no problems creating free tier compute instances but maybe it depends on region, I'm in Ashburn. Same I'm on a paid account, currently running a Kubernetes cluster for free is Phoenix. No issues converting from the free account and only once wasn't able to provision one of the ARM vms, and just provisioning in a different AZ worked. In Australia (and Singapore, soon) there is BinaryLane: https://www.binarylane.com.au/ (Disclosure: one of the co-founders, though I no longer work in the business.) Hetzner Cloud When I signed up and put in my credit card details, they immediately cancelled my order and completely deleted my account. I had to resort to moaning at them on twitter to get them to unblock me, which they did. Bit of a rubbish start to the journey though. Very strange, I've heard of multiple people having these sorts of issues with them, but it just worked when I signed up. Maybe they have some overzealous fraud detection going on? I've had a great experience so far with UpCloud. Almost a comparable suite of products to DO and their support team is very responsive. European based too which helps for some projects (GDPR etc) - though you can spin up instances in all the usual global regions. >Something has changed over there cough IPO cough I have been frustrated a bit with them as well because in the last few months, more often than not, a droplet hangs without any issues and they cannot tell us why. The only way is to reboot it. We were paying them over 5K/Month for some of our services and now I am moving everything to AWS EC2. Yes, I know that AWS can have its own issues but i have found their support to be better and so far more reliable infra than DO. I love DO for its simplicity and cost but lost trust recently due to those "droplets hanging suddenly" issues. That bites. At the same time, this is why the concept of "business continuity" exists. If you are running a hobby-level service, it may not be worth the trouble, but for anything where the continued existence of the service truly matters, you should have a plan in place of what to do when (not if) your host blows up and leaves you with nothing. That should include the ability to re-launch the code/database/server/whatever on a new host, and restore the latest data backup to that new host. I get that planning ahead for such things increases the work to launch a service - but anything short of it means you are not in control of your own destiny. You actually made me go look (I have a droplet as a Wireguard VPN, that I too use only rarely): Had trouble signing up for Digital Ocean, called up their support team, got a really bad response and decided I'll just blacklist them. If they were giving me trouble during sign up then what hope was there for support afterwards? Probably your server leaked, just move on. AWS lightsail is good enough. What leaked in this context ? Are you saying that is the only message you got? Maybe you got compromised? Have you tried reaching out to them? So go to any provider website, order bare metal server where proxmox installation will cover your requirements. Own VPS hosting is always better then cloud vps. Tell HN: Not your machines, not your servers Most might disagree but I actually think you have a point. I run most of my services off a raspberry pi and old MacBook and only use a VPS for WireGuard + port forwarding these days
I can't use it for getting to Streaming sites when abroad (they all blacklist DO's IP blocks), but it works for just about anything else and it's got really fat pipes. $ uptime
10:40:45 up 654 days, 11:32, 1 user, load average: 0.08, 0.02, 0.01