Wait until you hear what he did to h1b program!

The 100k fee on new applicants? drop in the ocean.

The h1b people spend (most) of their salary in the USA and pay US income taxes. Whereas overseas labor spend their salary over seas and pay no US taxes.

> You are not Google.

You think they are just running it for fun? It's literally non-Google customers who wanted this as was explained in the article

It's not hardcoded and you can increase it via flag.

There is a hard coded warning which says safety not guaranteed after 8GB. I have tried increasing this after a database has become full and it didn’t start. It’s definitely not a recovery strategy for a full etcd by itself, maybe as part of a way to eek out a little larger margin of safety.

This warning seems to be outdated. We had run etcd at much larger volumes without issues (at least without issues related to its size). Alibaba has been running 100G etcd clusters for a while now, probably others too

Thank you for the update

It's totally possible to run tens of thousands of QPS on etcd if your disks are NVMEs (or if you disable fdatasync which is not recommended). If you use kine+cockroachdb or tidb you can go even higher which is what I'm guessing is equivalent to their spanner setup.

I'm not sure why americans are apparently allergic to stucco vs expensive siding - it's just much better product and imho better look and no concerns with seams etc

If you want to realize that the USA is a republic, building will teach you quickly. Normatives and standards are localized.

Stucco is the default in the Southwest, especially in the temperate climates. Most lower density commercial nationally will use some kind of EIFS (think fast food establishment, drug store etc).

In places with more weather, it's not easy to get right and the residential crews aren't always abreast of EIFS that has well designed moisture management. I had to pull the entire failed stucco system off my current dwelling and replace it. There is no easy way to do that, and putting up a new one is also extremely labor intensive.

Stucco is not the default even here in NorCal where climate is as good as it gets... I don't think it's so much to do with "republic" but more with the fact that any given county permit office seems to be allowed to do its own thing. Evidence to that is even neighboring counties in my state have wildly different standards. Wrt to stucco though I think the technology is fairly standard and it's just the level of workmanship of contractors is so low that you might as well pay $$$ for relatively idiot proof product

The building code is legislated and enforced by the local elected governments, so, yes.

The classic Western stucco gig is to put down two layers of building paper over the cladding, styrofoam, chicken wire, a cementitious brown coat and one or more cementitious skim coats. This is really inadequate for anywhere with snow freeze/thaw cycles or enough rain and high humidity, but does fine in climates like central and southern CA, AZ, NV elevation dependent. You can improve this slightly by specifying different foam and skim product but to really deal with climate you are getting into manufacturer's product line, and it gets both technical and labor intensive where you are looking at an EIFS with proper moisture management.

I can’t speak for the whole (pretty large and varied) continent, but I once lived in a wet area that used stucco siding extensively in the 80s, and it became very susceptible to rot and mold.

Probably non zero number of companies use cloudfront and other cdns as fallback for cloudflare or running a blended cdn so not surprising to see other cdns hit with a thundering herd when cloudflare went down

Long time ago Google had a very similar incident where ddos protection system ingested a bad config and took everything down. Except it was auto resolved in like four minutes by an automatic rollback system before oncall was even able to do anything. Perhaps Cloudflare should invest in a system like that

Default podman network driver is just standard linux bridge

That's layer below - using air+pigeon as your PHY and paper as your MAC

TCP congestion control is performed by the end device based on mostly packet drop feedback. Middleboxes just drop traffic. Though sometimes there are techniques such as AQM (active queue management) where middleware router sends a special ECN (exlicit congenstion notification) bit in the packet but then still that relies on both endpoints to support and respect that notification.