You see cheaters banned in real time in Valorant, and the match is canceled, at least you used to.

Trolling is trying to piss off as many people as possible, using such a word achieved exactly that.

There's technically a way[1], but you'd have to do it every 6 months which is not great.

https://docs.github.com/en/communities/moderating-comments-a...

Yeah, that's actually what we've done on the Zig GitHub repository. However, it doesn't stop pushes to existing PRs, which isn't ideal; and, yes, it's quite hard to escape the conclusion that there being no "until I turn it back on" option is intentional.

It's completely intentional, and goes back to when GitHub was founded. GitHub was intended as a collaborative software development platform, not "look but don't touch".

I suppose you can fork a repository if you want to collaborate with others though. Reviewing pull requests and engaging with a community is a lot of work and has possible legal ramifications; in many cases it’s faster to just do things yourself. Some teams/companies deliberately refuse outside contributions for this reason.

You can close them and limit discussion to contributors I guess? Not ideal but at least they wouldn’t appear in the pull requests tab.

Alternatively you can use a bot or a GitHub Action to automatically change the description and title of the pull request to something like “[PRs are not allowed and deleted automatically]”. But yeah not a perfect solution either…

What's crazier is that a lot of it is not even JSON, so it's not being formatted at all.

And no lesson about single point of failure and centralization was learned that day.

Only true if your audience doesn't require Edge distribution, also if your Origin can handle the increased load and security issues, also if you don't use any advanced features (routing, edge compute...).

Aren't there other companies offering edge distribution?

Yes, and they come with different architectures, SLAs,... not so easy to pick one. Their features may not map 1:1, not so easy to implement a multi-edge solution.

Certainly. But I think that comes with the responsibilities, to be ready to switch a third party provider if necessary. Nothing is ever certain.

People are ready to switch to third party but such plan comes at cost.

Instant switch means really high cost (two contracts) and also maintaining and testing it regularly. Also means you are limited on advanced features and most likely stick to basic features, reducing your ROI.

Most people are ok with a switch that would take days to weeks, reimplementing basic stuff during initial migration, then iterating on more advanced features. You run at risk to be down for hours to days. The cost of 2N or 2N+1 vendors is just too high to justify it.

If your site is only hosted on one server and it catches fire, you can swiftly reinstall on a new server and change the IP your domain is pointing to, too... Still a single point of failure.

Bold of you to assume the service you use to manage your DNS was not also relying on Cloudflare just like you

Not if you’re using Workers/Pages!

But you didn't, so Cloudflare ended up being a single point of failure for half the internet.

If I had pointed my name servers somewhere else, then that of course would be the new single point of failure. You can't escape it, no matter how much hacker snark and down votes you have.

Almost like centralizing everything on a single service has consequences.

"Leave the billion dollar corporation alone!"

Nice Slashdot comment from 2004.

Found the "AWS certified cloud engineer".