Latency > 300ms - on simple health_check actix-web app

Love the platform… however…
I am running a tiny rust app in fly.io and I see my logs servicing the /health_check request in under 1ms however I get 300ms requests from syd to Melbourne which should be in double digit ping time territory … is this in the ball park for perf or am I dreaming of sub 50ms requests?

Can you please confirm:

Your app is in Sydney and you’re physically in Melbourne?

Are you using a custom domain that may get routed through a proxy first (often cloudflare does funky routing stuff that could take your request around the world before reaching its final destination).

Thats correct my App is in Sydney and I’m in Melbourne

This is me not you… if I use a vpn based in sydney I get 166 ms which is much closer to what I would expect

Can you send us a traceroute to your app along with your IP?

If you visit debug.fly.dev, it should show which region you reached too.

Traceroute when not on VPN from Melbourne to Sydney

traceroute -I pstack.fly.dev
traceroute to pstack.fly.dev (137.66.23.58), 64 hops max, 72 byte packets
1 192-168-1-1 (192.168.1.1) 4.375 ms 3.846 ms 3.947 ms
2 nme-sot-dry-bras21-lo20 (10.20.22.137) 8.161 ms 8.309 ms 6.584 ms
3 27-32-160-129.static.tpgi.com.au (27.32.160.129) 7.326 ms 8.533 ms 7.718 ms
4 per-apt-stg-crt1-be20 (203.219.57.14) 92.217 ms 91.512 ms 91.736 ms
5 sin-eqx-aye-int1-hu-0-0-0-20.tpg.com.au (203.219.107.114) 92.673 ms 91.137 ms 91.335 ms
6 * * *
7 * * *
8 137.66.23.58 (137.66.23.58) 227.301 ms 226.957 ms 226.627 ms

Trace route when on vpn ( sydney located VPN)
traceroute to pstack.fly.dev (137.66.23.58), 64 hops max, 72 byte packets
1 10.200.0.1 (10.200.0.1) 20.164 ms 18.696 ms 18.319 ms
2 139.99.131.253 (139.99.131.253) 19.660 ms 19.576 ms 19.724 ms
3 * * *
4 * * *
5 * * *
6 syd-sy2-bb2-a9.au.asia (103.5.14.218) 23.077 ms 20.135 ms 20.341 ms
7 syd-sy2-bb1-a9.au.asia (103.5.14.222) 19.870 ms 20.600 ms 22.652 ms
8 * * *
9 * * *
10 137.66.23.58 (137.66.23.58) 19.056 ms 20.021 ms 19.748 ms

thanks for the support !

Hi @scomp

Looking at your traceroute it seems that you’re being routed through Perth and Singapore which would explain the insane ping from Melbourne to Sydney.

Based off your traceroute I’m guessing your ISP is TPG, or a subsidiary of TPG, and doing some quick online searches it looks like they’ve had quite a few routing issues in the past with some happening this year.

It might be worth contacting TPG (or your ISP if you’re not with TPG directly) and asking them to investigate the issue.

What do you get when you go to debug.fly.dev when you’re both connected to a VPN and not connected to a VPN?

1 Like

I’m consulting with our upstream network provider, but it does work a lot better if pressure is applied on both sides :slight_smile: (your ISP and our network provider).

Will let your know if we find / can do anything. If we can you, should see an improvement in the next few days.