[yebyen]

That's $150/mo before you even launch a single worker tho.

I don't know about you, but I'd personally rather pay the $5 if I don't need a serious cluster yet. How much in actual absolute dollars do you think this is going to cost you, just in order to figure out exactly how it works, and what other AWS widgets you'll also need to pay for in order to use it effectively?

I am a human, basically looking for a stepping stone on some actual cloud to land on that is comparable to Minikube, and this sure isn't it.

[yebyen]

I'm pretty clearly in the minority with my opinion here, and I recognize that.

Amazon has made a substantial investment in Kubernetes here and that is nothing to shrug at. They are entitled (and justified) to recoup it. It's fair to say that K8S is not for lightweights, too.

AWS provides a fairly generous free tier and they are also justified to extract some payment from every user of their services that derives some non-negligible value. If you're using Kubernetes and you're not getting any value out of it, I think it's also fair to say that you're doing it wrong.

[thockingoog]

I serve a couple small websites from a very very small GKE cluster that costs well under $100/month. I don't worry about OS updates or Kubernetes upgrades. I don't worry about it going down.

I don't think my needs are that unique.

[dc_gregory]

Not unique, but certainly low value to Amazon in terms of expected revenue.

[user5994461]

You're not in the minority. It's well understood that there are amateurs trying to use the expensive clouds and shave off every penny, while the main customers are enterprise customers who couldn't notice a thousand dollars more or less.

It's just pointless for small users to argue about AWS costs. The service is not meant for them and it cannot fulfill their requirements.

[zzzcpan]

I believe the whole kubernetes is designed with the assumption that you are an enterprise customer.

[yebyen]

You're obviously no slouch, and I respect your opinion, but I think Amazon is all about the long tail and without speculating too much about what they are thinking, I'll say they might disagree. They are obviously smart to go for the big money dollars first, but everything I know about "Long Tail" says you're wrong.

I absolutely agree with everything you said, if you only added "at this time" to the end of that last sentence.

[user5994461]

What long tail do you mean? I would say the small tech companies with tens of developers and tens of servers are the long tail for AWS.

The service is too expensive and too complicated to be used by very small companies or individual users. There are also more appropriate competitors in that space.

[yebyen]

You told me you thought I'm not in the minority, and others have come forward to agree. I think we are the long tail.

It is exactly as you say, companies that don't want to spend more than they absolutely need to on infrastructure. Shaving pennies to save money. We are each too small to make any significant money on us, taken individually. That's what makes us the long tail.

Do you think in 6 months, you will be able to get a non-HA Kubernetes master on Amazon's free tier? Almost surely within 12 months. At that time, they will have addressed the long tail. Today's announcement is not for us. It's for large enterprise customers that settled on Kubernetes (and 57% of Kubernetes is already on AWS, so many of them are likely not new customers.) We're both right, from opposite perspectives.

[user5994461]

Sorry, there might be a misunderstanding. I meant that you were not in the minority, in terms of user count. There are people trying out AWS/Google, riding on the free tier, or just running a small site for the experience. People spending less than a hundred dollars a month could very well be the majority of the user count.

It doesn't mean that they bring any significant revenues or that the distribution follows a long tail. I would actually bet that amateur users are not forming a long tail.