Proxy cannot reach app & machines restarting too much

[lei1]

Hi,

We are currently experiencing an issue with our Fly.io application where the proxy is unable to reliably reach the app, and the service shows unstable behavior.

Specifically:

• The application often takes a very long time to load (sometimes several minutes or even longer)
• In some cases, the app gets stuck during initialization and does not respond
• The dashboard shows warnings such as “Proxy is having trouble reaching app” and “Some machines restarting too much”

Our app is deployed with multiple machines, and the issue persists even after restarts.

We would like to understand:

1. What could be causing the proxy connection issues in this case?
2. Why are the machines restarting frequently?
3. Is this related to resource limits, configuration, or regional deployment?
4. What steps do you recommend to stabilize the service and ensure consistent availability?

This issue is currently blocking our ability to properly test and validate the system, so we would greatly appreciate your guidance on resolving it.

Looking forward to your support.

Best regards

[halfer]

Have you tested your application locally in Docker?

Have you looked at machine logs from your Grafana instance? Either way, please post some example logs here.

Also check your memory and CPU graphs, and post them here. Your machines may be under-specified.

Is your app written with defensive error checking? What exactly happens if the app can’t reach the database?

There isn’t much information to go on in your question, but currently I would assume the app itself is unstable, possibly due to OOM issues.