Comments URL: https://news.ycombinator.com/item?id=48560859

Points: 5

# Comments: 0

When you're running hundreds of integration test suites per day in CI pipelines, the free tier is irrelevant. You need fast, deterministic, isolated environments that spin up and tear down in seconds, not real AWS calls that introduce network latency, eventual consistency flakiness, rate limits, and costs that compound with every merge request.

It'd be great to just use AWS but in practice it doesn't happen. Even if billing doesn't, limits + no notion of namespacing will hit you very quickly in CI. It's also not practical to give every dev AWS account, I did it with 200 people it was OK but always caused management pain. Free tier also don't cover organizations.

> they MUST learn that there are no hard spend limits, and the only way to actually learn it, is to be bitten by it as early as possible

This is a bizarre take. "The best way to learn fire safety is to get burned." You can understand AWS billing without treating surprise charges as a rite of passage.

And updated domain: https://mustafaakin.dev/posts/2016-02-08-writing-my-own-init...

https://duckdb.org/2025/03/12/duckdb-ui

I use every possible opportunity to say "Fuck Ubuntu Snaps"

Sure, we should cut waste, but compression exists for a reason. Dropping valuable observability data to save space is usually shortsighted.

And storage isn't the bottleneck it used to be. Tiered storage with S3 or similar backends is cheap and lets you keep full-fidelity data without breaking the budget.

Comments URL: https://news.ycombinator.com/item?id=43757666

Points: 2

# Comments: 0

Comments URL: https://news.ycombinator.com/item?id=43360789

Points: 6

# Comments: 0

I get that local disks are finite, yeah, but I think the core/memory/disk ratio would be good enough for most use cases, no? There are plenty of local disk instances with different ratios as well, so I think a good balance could be found. You could even use local hard disk ones with 20TB+ disks for implementing hot/cold storage.

Big kudos to the PlanetScale team, they're like, finally doing what makes sense. I mean, even AWS themselves don't run Elasticsearch on local disks! Imagine running ClickHouse, Cassandra, all of that on local disks.

I get that local disks are finite, yeah, but I think the core/memory/disk ratio would be good enough for most use cases, no? There are plenty of local disk instances with different ratios as well, so I think a good balance could be found. You could even use local hard disk ones with 20TB+ disks for implementing hot/cold storage.

Big kudos to the PlanetScale team, they're like, finally doing what makes sense. I mean, even AWS themselves don't run Elasticsearch on local disks! Imagine running ClickHouse, Cassandra, all of that on local disks.