Error: Failed to connect to Stomp!

[toyakoichi]

Our application executes the process of joining a lobby with the following steps:

1. Client A creates a lobby.
2. Another client B joins A's lobby.
3. B sends packets to A using P2PInterface.SendPacket. B continues sending packets at regular intervals until it receives a reply from A.
4. After receiving packets from B, A replies to B using SendPacket.

In this process, we occasionally encounter an issue where A fails to receive packets from B. The log on client B shows "Failed to connect to Stomp!".
Why does this error occur? Is there a way to prevent it?

---------------- (log) ----------------
2024-06-05T18:36:30.0414571+09:00 LogEOSMessaging(Error): Failed to connect to Stomp!
2024-06-05T18:36:31.4744445+09:00 LogEOSP2P(Warning): Waiting for PeerConnection task to complete before restarting connection.
2024-06-05T18:36:40.3242307+09:00 LogEOSP2P(Warning): Waiting for PeerConnection task to complete before restarting connection.
2024-06-05T18:40:38.2270204+09:00 LogEOS(Warning): Error response received from backend. ServiceName=[Presence], OperationName=[UnsetPresence], Url=[], HttpStatus=[404], ErrorCode=[errors.com.epicgames.social.presence.user_is_not_effectively_connected], NumericErrorCode=[24115], ErrorMessage=[Connection of facb4f298c6d40a781575f9cec1d2aa5 user not found in fghi4567BaAQtVx7S4zjsA9LIieXIgaT application and 448eca15031f4a94817cf9a59fcdd763 deployment], CorrId=[EOS-S4RIWUV3JESZvFBHIA9n7g-shyYq3uGPUqNGPWX40dGRg]
2024-06-05T18:40:38.2320058+09:00 LogEOS(Warning): Error response received from backend. ServiceName=[Presence], OperationName=[UnsetPresence], Url=[], HttpStatus=[404], ErrorCode=[errors.com.epicgames.social.presence.user_is_not_effectively_connected], NumericErrorCode=[24115], ErrorMessage=[Connection of facb4f298c6d40a781575f9cec1d2aa5 user not found in fghi4567BaAQtVx7S4zjsA9LIieXIgaT application and 448eca15031f4a94817cf9a59fcdd763 deployment], CorrId=[EOS-S4RIWUV3JESZvFBHIA9n7g-xPYeA1Sc5EWEy2FuR_So3Q]

[skeyll]

This is just speculation, but it seems that EOS need query NATType before connection.
The client tries p2p on EOS server, and at that time, it determines the connection type based on the queried NAT info. Therefore, if the connection request reach to the server before finishing query, it seems to fail. (The connect way is different between the client that queries or not.) I found that the initial connection became stable after waiting for the query result (NATType != NATType.None) before send a request.

[toyakoichi]

Thank you very much for the valuable advice.
Upon reviewing our project, we discovered that there are cases where communication starts before QueryNATType is completed.
We will make corrections regarding the points you mentioned and confirm.