@silverwizard Didn't #Vultr move to #DigitalOcean's model of requiring a credit card on file and monthly billing? Because of my income pattern, monthly billing is a no-go.
I'm sure most companies don't keep the full card info around, but instead are given a handle that enables them to charge the card each month without keeping dangerous info around. (Handle is my word for something I surmise exists but have no proof thereof.)
None of the articles I've seen mention the existence of such a thing, and therefore, don't say whether that information was also compromised.
I mention it because I haven't heard anyone talking about having to replace their cards or to cancel and restart monthly #DigitalOcean billing.
One last thing. Historically, when a company is breached, they say "the incident only affected a small subset of our users / customers", then that subset gets larger and larger over time. In some cases, the subset eventually comprises the entire user / customer base.
I would much rather pay for a year at a time, then about a month before the next payment is due, I get an e-mail reminder that payment is due by X date at Y time in Z timezone.
But if you’re in a place where that payment plan works for you, both of them were great places to host things over the years. I liked Linode a little more than I liked Digital Ocean, but your experience could be the opposite.
(1) #Twitter has millions of users. There is no #ActivityPub nor #OStatus implementation in which an instance hosted on a $5/mo #DigitalOcean / #Linode / #Vultr #VPS could handle the volume of a seamless connection with #Twitter. If they adopted AP OStatus, #Diaspora, or any other current open federation protocol, instances that didn't use firewall blocking would topple once the two userbases had sufficient interconnections (within a few hours or a few days after they started federating).
(2) Twitter's business model is to push ads disguised as tweets. If their users could escape those and still interact with all the same contacts, they would. I'm certain that Twitter's management know this. They also turn all links into tracking links, and sell access to media (images, video, audio) uploads of important news events to news organizations.
(3) Most Fediverse instances are financed out of the admin's pocket. Some have financial contributors, but nothing like Twitter's revenue. As the largest and best-financed instance, they would immediately have to start implementing modifications to make AP or other existing federation protocols useful to them, and those modifications would (as Mastodon's currently do) become unofficially mandatory in order to be compatible.
(4) This isn't the first time that Twitter has considered federation, though this may be the first time they openly discussed it. Back when Identica was still a happening place (during Twitter's fail-whale days), Twitter considered federating. They didn't do it then, and I honestly do not believe they will do it now.
(5) I'd say that Twitter's #BlueSky initiative is more meant to try to get bidirectional connections across #Facebook's moat and wall than it is to surround Twitter with a cloud of #Fediverse instances.
@feld That was a #Digital_Ocean change and one of the reasons I don't use them any more. I had multiple "droplets" paid annually (with a different prepaid card, but still prepaid) and suddenly I couldn't pay for them without giving them a regular card.
@1iceloops123 I’ve never had any site taken down over content, but if you anticipate posting things that would be subject to take-downs, you’ll have to ask someone else where to host it.
From what I’ve experienced, the $5-$10/month tier of #vultr or #digitalocean or #linode should be fine for a small instance. But again, that does not account for content-related take-downs.
I struggled to get Mastodon working with #DigitalOcean#Spaces + #CDN because the right .env.production file configuration settings eluded me. Here's what works for me:
That does mean that I can close my unused #DigitalOcean account, as this is a violation of their ToS. The DO account was unused for other reasons already. (Their resource limiter kept killing both MySQL and PostgreSQL database processes every few minutes, no matter how large the droplet, so what was the point of having the droplet?)
@moonman If you're on #Digital_Ocean, my experience is that something about their droplets' default configuration hindered both #MySQL / #MariaDB and #PostgeSQL performance. A smaller resource #VPS on #Linode runs them much better than DO.
.
I may try one on #Vultr soon.