Would love to know more about it, time for a new blog post/video 🙏

Currently running a bunch of web scrapers in k8s, only in one region for now, but need to expand to multiple regions soon but also don’t wanna be dependent on central APIs, so I’m in similar shoes as fly 👀

I'm asking because I'd assume it makes more sense to spread the machines across multiple physical hosts (even though they're most likely in the same DC or even rack) just so that in case one host fails, not all your machines for your app in that region are offline

Really awesome explanation! I've got a question though, if the workers make bids, does that mean the winning bid will always receive 100% of the workload of the bid? Let's say I want to scale to 3 machines in AMS, would all of them always end up on the same machine?

Oh dang, thought we already had it on our enterprise org 🙃 Joined the waitlist though hehe

I guess can’t really hurt, can just ignore it or remove it altogether if it’s ass, but might really be handy to find some things you overlook

Is it a new feature? Haven’t seen that one before

Looks like Graphana, how’s the data being sourced? Could it be that your instance itself reports the usage data and runs into the OOM error and then can’t report any updated metrics anymore so you won’t see the actual spike that triggers the error?

Absolutely no plan if that’s the case though lol