Hi Fly team,
I’m seeing very high latency for my API hosted on Fly.
App: `paytix-lu-api`
Custom domain: `api.paytix.lu`
Fly hostname: `paytix-lu-api.fly.dev`
Public IPv4: `66.51.126.67`
Primary region: `cdg`
Client location: Luxembourg / nearby EU
My machines are running in `cdg`
However, requests to https://api.paytix.lu/ are consistently very slow from Luxembourg:
curl total: ~930-1050ms
TCP connect: ~230-250ms
TLS complete: ~460-490ms
TTFB: ~930-1050ms
ping avg: ~226ms
Example curl result:
http_code=200
dns=0.002s
connect=0.244s
tls=0.484s
ttfb=0.948s
total=0.949s
remote=66.51.126.67
The response headers include a Fly request ID ending in nrt, for example:
fly-request-id: 01KWM57Z224HQZ9DFH0W4TDV9M-nrt
The route itself is trivial and returns a small JSON response, so this does not appear to be application/database latency.
Could you check whether the public IPv4 66.51.126.67 is being incorrectly routed through/near nrt instead of a European edge? Should I allocate a new IPv4, or can this routing be fixed on Fly’s side?
Thanks!