connection errors not propagated

[segfault-magnet]

It seems that if the committer experiences networking issues towards eth alloy will just log them and reattempt connection in the background.

not sure what happens with our network requests in the meantime, do they error out or block until the connection is reestablished?

Investigate so that the liveliness probe can be updated.

{"timestamp":"2024-10-03T11:30:35.142745Z","level":"ERROR","fields":{"message":"WS connection error","err":"WebSocket protocol error: Connection reset without closing handshake"},"target":"alloy_transport_ws::native","line_number":158}
{"timestamp":"2024-10-03T11:30:35.142822Z","level":"INFO","fields":{"message":"Reconnecting pubsub service backend."},"target":"alloy_pubsub::service","line_number":63}