Sometimes we have PU02 errors without any reason


Since we use we have sometimes a “PU02: Failed to complete HTTP request to instance”. I want to say without reason because the traffic is low at this time. We have a ROR application with puma server on front 6 workers / 6 threads. The client of our backend have sometimes 502 error and it matches with PU02 in our logs. What can produce this issue? we have to less information in logs…

They added an errors page to their docs that explains each code: error codes and troubleshooting · Fly Docs

Some errors (like the one you listed) doesn’t have a lot of information unfortunately.

Yes I already seen in the doc.

When did you start seeing this error? I have an unrelated error regarding dns hostname not being reachable when my instance boots up, it started around June 15th… happens 3-5 times each day.

It seems that it was because of a bad configuration in the fly.toml
Indeed by mistake, in the [http_service] section we put processes = [“app”, “worker”], worker mustn’t have to be here :slight_smile: