Strike48 Labs Strike48 Labs

StrikeHub: Connectivity

This guide covers common connectivity issues between StrikeHub and its connector applications.

Connector shows "Offline" but it's running

Symptom: A connector process is running but StrikeHub shows it as offline.

Possible causes:

  • Socket path mismatch — The Unix socket path in the connector doesn't match what's in connectors.toml.
  • Health endpoint not responding — The connector must serve GET /health and return 200 OK.

Debugging:

# Check if the Unix socket exists
ls -la /tmp/strikehub-kubestudio.sock

# Test health over Unix socket
curl --unix-socket /tmp/strikehub-kubestudio.sock http://localhost/health

Tip: Check ~/.config/strikehub/connectors.toml to verify the socket path for each connector.

Socket path mismatch

Symptom: StrikeHub logs show connection refused for a specific connector.

Possible causes:

  • The connector binary was started with a custom socket path that doesn't match the config.
  • A typo in the socket filename.

Debugging:

# List all StrikeHub sockets
ls /tmp/strikehub-*.sock

# Check the configured path
grep socket ~/.config/strikehub/connectors.toml

Ensure the socket path in the config exactly matches the one the connector creates.

WebSocket connection drops

Symptom: The connector UI loads but becomes unresponsive, or live updates stop.

Possible causes:

  • WsRelay crash — The WebSocket relay bridge has exited. Check StrikeHub logs for ws_relay errors.
  • Network interruption — Brief connectivity loss between the connector process and StrikeHub.
  • Connector crash — The connector process itself has exited. Check the sidebar status indicator.

Debugging:

# Run with WebSocket tracing
RUST_LOG=sh_core::ws_relay=debug cargo run --features desktop

# Check if the relay port is still listening
lsof -i :9090

Caution: If the WsRelay repeatedly crashes, check for port conflicts on the relay port (default 9090).

Stale socket files

Symptom: StrikeHub fails to start or connectors can't bind their socket.

Possible causes:

  • StrikeHub or a connector exited uncleanly, leaving orphaned socket files in /tmp/.

Debugging:

# Remove stale sockets
rm /tmp/strikehub-*.sock

# Verify no process is using the socket
lsof /tmp/strikehub-kubestudio.sock

StrikeHub cleans up stale sockets on startup, but manual removal may be needed if the cleanup itself fails.