#openclaw gateway status shows successful probe, but openclaw gateway probe fails

1 messages · Page 1 of 1 (latest)

opaque valley
#

status shows:
Runtime: running (pid 12126, state active, sub running, last exit 0, reason 0)
RPC probe: ok

Listening: 127.0.0.1:18789

...but probe shows:
Gateway Status
Reachable: no
Probe budget: 3000ms

Discovery (this machine)
Found 0 gateways via Bonjour (local.)
Tip: if the gateway is remote, mDNS won’t cross networks; use Wide-Area Bonjour (split DNS) or SSH tunnels.

Targets
Local loopback ws://127.0.0.1:18789
Connect: failed - timeout

craggy vineBOT
#

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> âś… Mark Solution

alpine bluffBOT
#

That combo (“openclaw gateway status says RPC probe ok” but openclaw gateway probe times out to ws://127.0.0.1:18789) is most often scheme mismatch:

  • gateway status computes the probe URL as ws:// vs wss:// based on gateway.tls.enabled
  • gateway probe currently always adds the local target as ws://127.0.0.1:<port> (even if TLS is on), so if your gateway is actually TLS-enabled, the probe’s loopback target will just hang/timeout.

Can you run these and paste output?

  1. Version (helps spot known CLI quirks):
openclaw --version
  1. Confirm TLS:
openclaw config get gateway.tls
openclaw config get gateway.tls.enabled
#
  1. If TLS is enabled, try probing with wss explicitly:
openclaw gateway probe --url wss://127.0.0.1:18789

If that works, you’ve basically confirmed it’s just the gateway probe localLoopback URL being ws:// when your gateway is wss://. If you want, file a bug with those outputs: https://github.com/openclaw/openclaw/issues

If TLS is not enabled and it still times out, the next suspects are “you’re running the probe from a different network namespace (Docker/WSL)” or a stuck listener — but check the TLS bit first because it matches the exact symptom.

opaque valley
#

Config path not found: gateway.tls
Config path not found: gateway.tls.enabled

alpine bluffBOT
#

Config path not found: gateway.tls… just means you don’t have those keys set in your config (so TLS isn’t the explanation here).

What’s going on is simpler (and a bit annoying):

  • openclaw gateway status does its RPC probe with a very generous default timeout (10s).
  • openclaw gateway probe has a hard per-target budget, and for Local loopback it’s capped at 800ms. If the gateway doesn’t complete the WS “hello” handshake inside that window (CPU spike, GC pause, IO stall, busy startup, etc.), it reports:
    Connect: failed - timeout
    even though the gateway is actually fine (as your gateway status already proved).

Two useful next checks:

  1. Please paste:
openclaw --version
  1. Try the probe again after the gateway has been idle for ~30s. If it still always times out, that usually means the machine is pausing hard (CPU pegged / VM contention / heavy IO). To confirm that, run: