Open Source Durable Objects for Postgres | Hacker News Reader

Open Source Durable Objects for Postgres

solidobjects.dev

41 pointscardmagic24 comments

> Pricing you cannot predict. Every operation is metered. One runaway alarm loop billed a pre-launch developer $34,000 in eight days, with no users and no warning.

It's amazing that we still lack the ability to set a hard cost cap on a lot of cloud services/API, as the consumer you usually have a rough idea of what something is going to cost (I say rough because pricing out AWS is always fun) and an idea of what the lower bound "oh shit that isn't right" limit should be, so why can we not set that.

You shouldn't be getting an unexpected $34K bill (unless it's within variance and you are running a very large system).

Cynically the opaqueness of this stuff has to be by design/intention at this point.

Especially when Google and co tells you that they can bill by the second.

I do not own a hyperscaler account as a private person and will never do, the financial risk is too high even IF the support will swift it away.

It is perhaps more accurate to say that they can meter by the second, but the billing calculation is a trailing activity.

>as the consumer

This is where this discrepancy occurs. AWS (etc.) aren't targeting "consumers," they're targeting enterprises. Enterprises want to use these services with minimal interference and overhead. They'll build out whatever features they need as they need them. It's the same thing with any other utility service: a factory isn't interested in having the electric company or water company shut off their services if their usage spikes. They'll manage their usage (and potential overages) in whatever way is specifically suitable for them.

Consumers have plenty of options to prevent such situations. Hell, AWS even offers Lightsail which is the consumer-friendly version of their services which is designed to be simple with a flat rate cost (including options of object storage).

If a consumer decides to tap into enterprise-grade services and ends up getting burned as a result, then I'm blaming the consumer rather than the provider of those services. These dynamics are all pretty clear and easy to understand (and even mitigate, with enough effort), so I don't see how you can claim they are "opaque" unless you're the kind of consumer who feels every company should dumb down their products to cater to the lowest common denominator.

What does a "hard cap" do then? Delete all your resources? Because even stopped instances or unused volumes cost you.

Although I admit, there should be spend warnings.

If what you're doing makes money, the last thing you want is to stop doing the thing making you money. If, instead, it costs you money, don't wait until a usage surge to consider how to handle that – set it up now.

You also don't need any latency in the loop or moving parts that could wrongly interfere with the reliability of making that money.

Fortunately, the last thing this is is "opaque". You have all the observability you need to do what you want to do with unexpected usage.

What AWS (and every CSP) gives you is all the instrumentation and out of band tools to implement your own angel of death, to API kill a runaway thing if that's what you want to do.

You can also find countless TF and CDK patterns for it if you aren't sure how.

But please, no, do not ask to make critical systems subject to the whims of some billing decision enforcer.

The "why did they cut me off" complaints here far outnumber the "runaway bill" complaints.

I don't know if it is ai generated or at the opposite hand written a little bit fast but I find this article very hard to read despite the concepts in it being simple.

I couldn't parse "Getting that model has meant taking on something else."

The article is 1000% ai slop

Hard disagree there. I still don't know what the article is about if I'm being honest, but it doesn't bear the markings of being generated.

you need to tune your slopmeter, friend

AI, but I'm not sure about "slop". There is actual content in there, for all it's written by an LLM.

Yeah, i just call anything thats obviously been generated by AI "slop".

In my mind, slop is when the output is longer than the context provided by the user -- there's text, but there's no content.

Where LLMs can be really actually quite useful, is situations where you have a lot of context and want to distil it. That's what I think happened here: the author has a lot of information to convey, and the LLM-generated output is relatively information-dense.

Unfortunately there's no objective measure for information density of an arbitrary piece of text.

im not against using AI for writing, learning, etc... But if you publish/share writing that is full of LLM-isms, its slop. Full stop.

The criticism of Cloudflare here is reasonable, I think. I really like the CF building blocks - they're neat and conceptually useful and quite unusual. I think they could have a better developer experience though (though it probably is better than when I last tried it), and I also think they could have a little more confidence in their overall platform's usefulness and open up more of their model as open source. For example, I'd love to run my CF workers in CI locally without creating infrastructure (again, this may have changed since I last tried it).

wrangler / workerd enables local CI

I might have to have another look. It always felt a bit magical, but maybe it's not so magical any more.

Please at least write the announcement post by hand.

Yeah it's hard to read, so dry and AI

Why chose the Solid- line of naming, considering that's the rails provided naming. I makes it look like this is official.

Author here.

celld and workerd give you Durable Objects if you run another daemon. This is a library your app loads, with state stored in the SQL database you already have (Postgres, MySQL and SQLite supported). No daemon required, no new infrastructure and monitoring, no broker, and no vendor account.

Pre-1.0, MIT, TypeScript, in browser (with WASM SQLite), and Ruby too. Happy to answer anything.

This looms nice and I starred but you just want actors and there are other options for that you might want to compare, I like rika actors, or temporal (not actors exactly but durable state and workflows). This is a space where a few very similar technologies after looked from different lenses sell themselves widely differently. So if you know what you want and the names of the things you are searching for you can get a thing that looks 90% like te rest, sorry a bit rambly. But yeah it would be nice to add some comparisons against these technologies as well.

Please don't submit AI-generated articles.

Reply on news.ycombinator.com