Resolved.. not running optimized.. Performance not matching up. Router is 200 ms slower on average. Not Sending APQ to Downstream Services

[golfingal72]

I am working with the new router and comparing performance and it's slower than the old gateway, even though documentation says it's much faster????

I am running locally all the processes in the graph as well as the gateway vs. the router. I de-activated my gzip compression that i have running in the old gateway to try to do an apples to apples comparison... I ran the same query which is large which has a large payload over and over again so ensure all caches are warmed up... on average the new router is about 200 ms slower...
I have to verify that the persisted query is getting sent to the backend services... will follow up

[golfingal72]

Are you sending the APQ to the downstream services?

[golfingal72]

it appears the router is NOT sending the APQ downstream... so this appears to be the performance issue

[o0Ignition0o]

Hey @golfingal72,

APQ is not sent downstream, we do have an issue tracking this here.

[golfingal72]

Does it matter if i am running it as a cargo run command?

One of my peers suggested, it is faster to run by making a release executable?

[bnjjj]

Yes indeed you need to make a cargo build --release to have an optimized binary, otherwise you will have bad performances

[golfingal72]

That indeed was the PROBLEM... I am new to Rust and didn't understand that ... thanks a bunch

[o0Ignition0o]

That's great news!
What kind of numbers are you seeing now?

[golfingal72]

i see it about 20 ms faster on my localhost... haven't done extensive testing yet