That being said, and I can’t put my finger on why necessarily, it sometimes feels a bit like Heroku - the thing you use before you “graduate” to just using one of the major cloud providers.
That being said, and I can’t put my finger on why necessarily, it sometimes feels a bit like Heroku - the thing you use before you “graduate” to just using one of the major cloud providers.
There is a difference between being cheap and inexpensive. DO is inexpensive, but high quality.
But support for more complex backend infrastructures targeted by the cloud providers, with network segmentation and load balancers and autoscaling clusters, has only started to grow in recently. There is nothing approaching any cloud provider's IAM.
IAM is a big no though.
Investors went into DO thinking of a AWS/GCP/Azure unicorn, but after 6 years it appears to be a small rainbow pony than an almighty unicorn. This restructuring appears to be a realization of what happened with that vision and how it panned out, I may be wrong though.
I personally love DO, I don't want to deal with AWS complexity for basic needs. Minimalism and simplicity certainly has value.
Definitely a skill floor/ceiling thing going on.
At that point, all I need from the hosting provider is a server that stays up with a network that stays up. And contacts to skip bullshit triage when the network is broken and their monitoring doesn't show it.
Conversely, I also feel like Heroku could round out their offering just a bit more (ex: built in robust monitoring and alerting, better deploy notifications and healthchecks, a simple custom metrics system) and really lock in that "premium basic" tier that would be perfect for most small to medium businesses.
I've always enjoyed Heroku, but feel like I'm paying a premium because I don't know enough to do "real" cloud work.
"If I were a more experienced developer, I could run this on AWS for pennies on the dollar".
Honestly, I don't know where that sentiment comes from exactly. It's at least partially coupled with the fact that all of add-ons you would need to pay for to run a small business site add up pretty quickly, from what I can tell.
There's some friction with AWS (e.g. SSL termination) that Heroku makes incredibly easy, so I usually don't migrate until it's clear that a service is going to continue to exist in the medium term - otherwise the investment isn't worth it.
Didn't mean to imply that was a widespread sentiment, of course! Just my personal one.
Agree. But I think part of the plan with Amazon et al is that those who RTFM are locked in to the platform and therefore will be less likely to switch.
AWS/etc heavily reward you for investing time/money/etc but can be a bit more difficult to get started with without investing resources immediately, whereas DO is very easy to get up and running on but can be difficult when you encounter anything more complex.
It's kinda nice because you're just putting up VMs and there's not the type of vendor lockin you get with AWS/Azure/Google. That being said, DO is obviously trying to compete on that scale now. It has managed databases to compete with RDS, load balancers, and even managed k8s.
They want to be a real AWS alternative, but when you start building around these components, you get locked in. If they're laying off people, it could adversely affect startups thinking of building around their services.
And for people who think "Well I can just use another terraform provider" .. it really is not that simple at all. AWS/Azure/Google/DO are all very different. They have vastly different terraform providers/modules and you're pretty much writing an entirely new setup per each provider because of they way they handle firewalls, security, IPs, inbound-outbound, etc. The OpenStack API or any type of real standard has yet to emerge that branches all these offering. If you want "cloud" hosting (and most do because managing your own Postgres/MySQL clusters with backups and failover is a fucking bitch; especially if you're just starting out and want to get going fast), you need to realize you might need to be locked in very early in the game.
They can't. Too small to grow big.
When I worked at DigitalOcean all we could do was shake our head and laugh at how blatantly hard Vultr tried to copy what we did.
An interesting thread on the Linode forum when DO launched: https://www.linode.com/community/questions/8303/new-linode-c...
Today, Vultr offers similar service, at a lower price point, without the risk that your entire account will be suspended for thinking unapproved thoughts. This is a pretty good selling point to anyone who isn't an apparatchik of modern political and social dogma.
https://battlepenguin.com/politics/the-new-era-of-corporate-...
Free speech maximalism as a cover for raging racism (and sexism, and homophobia, etc) is what's disgusting.