can't find how to configure machine restart / new machine creation policy on health check fail

Yesterday I had an app that started giving a lot of these errors

could not make HTTP request to instance: connection error: timed out

My app was running two instances in the sin region. Both of them started giving this error and the health checks were also failing.

I still haven’t been able to figure out the cause, but let’s assume that something went wrong (either fly’s proxy was not able to route or the app was not accepting connections).

My question is, how can I tell fly to create new machines automatically to make sure I have 2 healthy instances (since my scale count was 2). I can see a restart_limit flag but that seems to be for v1 apps.

What is the default behaviour expected to be and how can I do some configuration here. I am using a fly.toml file.

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.