Unreliable response times in EU for free plan

[furkankly]

Since yesterday, I’ve been experiencing unreliable response times in ams, mad and cdg. I tried all these 3 regions because the server has been responding at times, but hanging most of the time. No errors in monitoring, everything passes with flyctl doctor. I set up a healthcheck and it has been passing and failing intermittently as well.

Is this because Im using a free plan? I tried to scale it using flyctl scale count 5 but it downscaled automatically. But I think the issue is not related to this because I had no issues last month in cdg using a free plan.

Also, when the app was responding, I realized there were a few requests getting 502, happened to be only the favicon ones which is kind of odd. Its a Rust app serving a React spa:

[http_service]
internal_port = 8080
force_https = true
auto_stop_machines = true
auto_start_machines = true
min_machines_running = 0
processes = ["app"]

and all I do is building the app and exposing 8080 in my Dockerfile.

[holden]

I’m not on the free plan. But I’ve also had similar issues that started in the last couple of days where Fly seems unreliable all of a sudden. We’ve had our app up for probably 6-months without issue and haven’t made any recent changes that could explain this, but the last couple of days I’ve found myself scaling up, increasing ram, restarting, incantations, you name and it will eventually seem ok for a few hours then the next day I’ll need to mess with it again. All without any useful errors.

[furkankly]

Correction: I deployed it on AWS and came across with the same issues. Im guessing its related to how I setup my server using axum.