Current Behavior
The panel occasionally shows "Error connecting to node!" for a node that is actually online and responding normally.
This happens even when:
- Manual
curl requests to the node's Wings API succeed instantly (<50 ms)
- The node's DNS resolves consistently
- Wings is confirmed listening on the configured port
- 6 parallel manual
curl requests to the same node all succeed without issue
The failure surfaces as a fast ConnectionException ("Couldn't connect to server ... after ~30 ms"), not a timeout. This suggests a transient TCP-level failure (e.g. an occasional refused connection) rather than a systemic network or configuration problem.
Proposed fix: add ->retry(2, 200) to the request in DaemonSystemRepository::getSystemInformation(). I applied this on my live 2-node setup and the node list no longer intermittently shows nodes as unreachable. Happy to open a PR if this approach is acceptable, or to adjust it (e.g. retry only on ConnectionException).
Expected Behavior
A single transient connection failure should not immediately mark a healthy node as unreachable in the UI.
Steps to Reproduce
- Set up a panel with 2 nodes connected over a low-latency private network.
- Open the node list / node overview page repeatedly.
- Occasionally one of the nodes is shown as unreachable ("Error connecting to node!"), while Wings is up and reachable via
curl at the same moment.
Panel Version
1.0.0-beta38
Wings Version
1.0.0-beta29
Games and/or Eggs Affected
No response
Docker Image
No response
Error Logs
Is there an existing issue for this?
Current Behavior
The panel occasionally shows "Error connecting to node!" for a node that is actually online and responding normally.
This happens even when:
curlrequests to the node's Wings API succeed instantly (<50 ms)curlrequests to the same node all succeed without issueThe failure surfaces as a fast
ConnectionException("Couldn't connect to server ... after ~30 ms"), not a timeout. This suggests a transient TCP-level failure (e.g. an occasional refused connection) rather than a systemic network or configuration problem.Proposed fix: add
->retry(2, 200)to the request inDaemonSystemRepository::getSystemInformation(). I applied this on my live 2-node setup and the node list no longer intermittently shows nodes as unreachable. Happy to open a PR if this approach is acceptable, or to adjust it (e.g. retry only onConnectionException).Expected Behavior
A single transient connection failure should not immediately mark a healthy node as unreachable in the UI.
Steps to Reproduce
curlat the same moment.Panel Version
1.0.0-beta38
Wings Version
1.0.0-beta29
Games and/or Eggs Affected
No response
Docker Image
No response
Error Logs
Is there an existing issue for this?