admiralpatrick

Has anyone ever tested those red hats for lead content?

That's a good point. I'll have to check the default values, but on my own instance, I have very conservative limits in place, and it hasn't proven to be an issue (so far?).

Unless it's changed since I wrote the online docs for Tessreact, the modlog is part of the "Messages" rate limit bucket which is/was something of a catch-all for endpoints that didn't fit elsewhere. Even in the default config, that bucket is the most permissive due to that.

I've been daily-driving my dev version with this feature enabled for a few days, and it hasn't been an issue so far (it only does a modlog lookup if a comment is removed, so not on every comment in the tree). It's also per IP, so unless a lot of people are behind the same public IP, I don't think it's going to pose an issue. I'd have to double check, but I think the most comments it loads in a batch is close to 100, so unless every comment has been removed, that would be the worst-case number of modlog fetches.

So it looks like I'll definitely want to make this feature toggleable even if it does end up defaulting to 'on'.

I did test a few posts with a lot of removed comments (both on my instance and Lemmy World), and the overhead wasn't terrible since scoping the modlog to just the comment ID is pretty lightweight. Since HTTP/2 is pretty common (which can re-use connections), there's not overhead of additional TLS handshakes slowing things down.

From a network traffic standpoint, Gzip compressed JSON is pretty negligible in the grand scheme of things.

Put them in the body. I'll downvote every archive link used the post url because it obfuscates the source.

Yes, those are called "Opinions / Editorials / Letters to the Editor" et al. Both News and WorldNews have rules against opinion articles as well.