"could not wake up machine due to a timeout requesting from the machines API" after the app has already un-suspended

I see the following logs for my app that strike me as weird:

2025-10-14T11:34:11Z app[d8917e1ae64178] sin [info]Virtual machine has been suspended

2025-10-14T13:22:17Z proxy[d8917e1ae64178] sin [info]Starting machine

2025-10-14T13:22:17Z app[d8917e1ae64178] sin [info]2025-10-14T13:22:17.917676512 [01K7CAGXG7E0TN8C2PC5AG3TJ3:main] Running Firecracker v1.12.1

2025-10-14T13:22:17Z app[d8917e1ae64178] sin [info]2025-10-14T13:22:17.917812324 [01K7CAGXG7E0TN8C2PC5AG3TJ3:main] Listening on API socket (“/fc.sock”).

2025-10-14T13:22:17Z app[d8917e1ae64178] sin [info]2025-10-14T13:22:17.918125850 [01K7CAGXG7E0TN8C2PC5AG3TJ3:fc_api] API server started.

2025-10-14T13:22:17Z app[d8917e1ae64178] sin [info]2025-10-14T13:22:17.919937369 [01K7CAGXG7E0TN8C2PC5AG3TJ3:fc_api] The API server received a Get request on “/”.

2025-10-14T13:22:17Z app[d8917e1ae64178] sin [info]2025-10-14T13:22:17.919968830 [01K7CAGXG7E0TN8C2PC5AG3TJ3:fc_api] The request was executed successfully. Status code: 200 OK.

2025-10-14T13:22:17Z app[d8917e1ae64178] sin [info]2025-10-14T13:22:17.920421067 [01K7CAGXG7E0TN8C2PC5AG3TJ3:fc_api] The API server received a Put request on “/logger” with body “{\“log_path\”:\“logs.fifo\”,\“level\”:\“info\”}”.

2025-10-14T13:22:22Z proxy[d8917e1ae64178] sin [error][PM03] could not wake up machine due to a timeout requesting from the machines API

2025-10-14T13:22:23Z proxy[d8917e1ae64178] sin [info]Starting machine

2025-10-14T13:22:23Z proxy[d8917e1ae64178] sin [error][PM01] machines API returned an error: “machine still attempting to start”

2025-10-14T13:22:26Z proxy[d8917e1ae64178] sin [info]Starting machine

2025-10-14T13:22:26Z proxy[d8917e1ae64178] sin [error][PM01] machines API returned an error: “machine still attempting to start”

2025-10-14T13:22:35Z runner[d8917e1ae64178] sin [info]Machine started in 17.316s

2025-10-14T13:23:03Z app[d8917e1ae64178] sin [info]{“level”:“info”,“time”:“2025-10-14T13:23:03Z”,“message”:“gov:intercept routed%!(EXTRA map[string]interface {}=map[cid:6ea28e6d-0428-438f-8944-5c96977de590 model_after:openrouter/google/gemma-2-9b-it:free model_before:google/gemma-2-9b-it:free])”}

It seems like already after the machine has been un-suspended and was replying something, Fly Proxy attempted to re-start it again.

Note (if that matters): this wake-up was caused by a request via external proxy (…fly.dev), not .flycast. And that request happened at around 1760448133 (2025-10-14T13:22:13Z), which is 3-4 seconds before any visible signs in the logs above.

Hi… Three other fellow users mentioned very similar anomalies recently:

https://community.fly.io/t/another-one-for-persistent-machine-still-attempting-to-start-errors/26064

There hasn’t been a resolution yet, that I know of, but here’s a work-around that was recommended back then:

Hope this helps a little!

Please see Fixed: unreasonably slow resumes of suspended Machines

1 Like

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