My app didn't go down yesterday, this was a downtime related to internal API and some specific regions.

It's free until it isn't anymore

Egress was pretty much always unmetered until AWS came along. And it still is with most providers other then the usual suspects (AWS,azure, Google cloud)

Right, Cloudflare is operating in good faith, and while the deal could change (because business is business and sometimes the offering has to change), one should ensure they have other options available for egress from origin. That's not a Cloudflare issue, that's a vendor risk management issue. You should always have alternatives planned if a vendor's specific offering at a price point is a material component of your business costs.

unmetered != unlimited

There might not be precise measurement system which bills you by the byte, but many provides monitor (or periodically sample) bandwidth utilization, so and using too much data will catch someone's attention.

Example from Hetzer, which claims to provide "unlimited" traffic. https://lowendtalk.com/discussion/180504/hetzner-traffic-use...

> We will routinely check your traffic use. If we continue to see such a high amount, we may choose to cancel your contract.

That’s wonky. If their normal servers include 20TB, why would they consider 250TB on their ‘unlimited’ servers too much. Makes you wonder who they think opts for unlimited servers. If 20TB was enough they clearly wouldn’t.

And then later add a script that downloads the binary from Github releases. Doesn't improve the situation with curl script haters

I feel like the assumption is that GitHub would be more proactive about stopping malware being distributed from their platform.

Dive doesn't have a JSON output, I had to use the internal API to do it.

What do you mean? It's literally there:

  $ dive -j /dev/stdout alpine:edge
  Image Source: docker://alpine:edge
  Fetching image... (this can take a while for large images)
  Analyzing image...
  Exporting image to '/dev/stdout'...
  {
    "layer": [
      {
        "index": 0,
        "id": "blobs",
        "digestId": "sha256:8980900bae6f916f9dc1158ed71ee72566d7ad3f47f63fe19ba3ab7a4ef1a432",
        "sizeBytes": 7738376,
        "command": "#(nop) ADD file:5ba6edac87fa0feb475d0ddae7299210c6f2b69714784771ca86fcb6407ae8fc in / "
      }
    ],
    "image": {
      "sizeBytes": 7738376,
      "inefficientBytes": 0,
      "efficiencyScore": 1,
      "fileReference": []
    }
  }

It's a treemap graph, frontend people use it all the time to analyze a website javascript bundle size, I created this so Docker people can make smaller images more easily

Finally switched to Nvidia and already adding great value

Another pain point for me: no syntax highlighted code blocks

I see a lot of people using GitHub gists to embed code blocks for this reason, this means that writing any article code related is a pain

Edit: i think it's available now but definitely too late

Yes. We added them recently.

I do E2E tests for the core features against a staging database and url.

This way i don't waste time with unit tests that quickly get old and no one wants to maintain and run

If you want to start a blog quickly check out Notaku [0], it can create a blog website using Notion to manage content

[0] https://notaku.so/product/blog

The hard part of building a CDN is scaling it. The best approach imo is to use fly.io to host an anycast IP (with horizontal scaling) and store cache files on disk

Fly.io also has a Grafana dashboard built in for your machines

Agree, Fly.io is great for such usecases. Is there any CDN/Proxy solution or guide available for fly?