I’ve put in a PR so you can see what region the test is running in.
This is not likely to be an issue with the docker image or linux distribution. It looks like an issue in DFW. Most other regions get sub-10 (and usually sub-5) millisecond latencies using your reproduction code which we launched in a few places.
We’ll be looking a this more closely today and keeping you updated.
This should be resolved in DFW. We’ve noticed other regions where it’s much slower than it should be. We’re working on them now.
We’re running on multiple providers (Packet is our main one), but not on AWS. We’re running powerful bare-metal servers and we’re using anycast to globally route requests to the closest (generally speaking) server. Anycast is not available through AWS unfortunately, not many providers offer it (especially not “bring your own IP space”). As far as software goes: Linux with a fairly recent kernel, we built our own Rust-based reverse-proxy which feeds off of Hashicorp’s consul! Every app runs in a firecracker microvm, scheduled with Nomad.
Edit: added more details about what Fly runs on!
Oh yeah that’s what I’m talking about!!!
Redis latency test: Region: dfw 2ms 1ms 0ms 1ms 0ms 1ms 1ms 0ms 1ms 0ms
You’re the best!!!