OpenClaw Deployment Troubleshooting Checklist (2026): Gateway Fails, Telegram Bot Silent, Chrome Relay Not Connecting

If you just deployed OpenClaw, most incidents fall into three buckets:

1. openclaw gateway start fails
2. Telegram bot is online but does not reply
3. Chrome Browser Relay shows red ❗ or times out

This is a command-first troubleshooting runbook: stop the bleeding, locate the root cause, then stabilize.

0) Run a unified health check first

openclaw status
openclaw gateway status
openclaw channels list
openclaw doctor
openclaw logs --follow

What to verify:

• gateway status is running
• channels list includes telegram and it is enabled
• Logs show a consistent error pattern (409 / EADDRINUSE / timeout)

1) Gateway fails to start: fix by error signature

A. Error: Gateway start blocked: set gateway.mode=local

openclaw config set gateway.mode local
openclaw gateway restart
openclaw gateway status

B. Error: EADDRINUSE (port conflict)

sudo lsof -i :18789
# or
sudo ss -tlnp | grep 18789

openclaw gateway stop
kill <PID_USING_18789>
openclaw gateway start

C. Error: refusing to bind gateway ... without auth

You bound gateway.bind to a non-loopback address without auth.

# Recommended for single-host deployment
openclaw config set gateway.bind "127.0.0.1"
openclaw gateway restart

If you need remote exposure:

openclaw config set gateway.auth.token "YOUR_STRONG_TOKEN"
openclaw gateway restart

D. Service keeps crashing under systemd

journalctl --user -u openclaw-gateway --no-pager -n 80
which node

Check these first:

• Node version meets OpenClaw requirement (typically Node 22+)
• ~/.openclaw/.env has required API keys
• Service user can read/write OpenClaw config/data paths

2) Telegram bot silent: resolve 409 first, then token/webhook

A. Error: 409 Conflict: terminated by other getUpdates request

This means the same bot token is being polled by multiple instances.

openclaw gateway stop
ps -ef | grep -i openclaw | grep -v grep
# after cleanup, run exactly one instance
openclaw gateway start

If you run OpenClaw on both laptop and VPS with the same token, keep only one online.

B. Verify bot token validity

curl "https://api.telegram.org/botTOKEN/getMe"

If you get 401, regenerate token in BotFather, update config, restart gateway.

C. Remove leftover webhook (required for polling mode)

curl "https://api.telegram.org/botTOKEN/getWebhookInfo"
curl "https://api.telegram.org/botTOKEN/deleteWebhook"
openclaw gateway restart

D. Pairing policy blocks replies (dmPolicy=pairing)

openclaw pairing list telegram
openclaw pairing approve telegram <CODE>

3) Chrome Relay not connecting: check service → port → tab attach

A. Confirm browser control service is reachable

openclaw gateway status
curl http://localhost:18792/health
ss -tlnp | grep 18792

B. Extension red ❗ or timeout

Typical causes:

• Gateway not running
• Port 18792 not reachable
• Local browser trying to connect to remote gateway without tunnel

For remote hosts, create local SSH tunnel:

ssh -L 18792:localhost:18792 user@your-server

C. Error: no tab attached

This is not a crash — it means you did not attach a tab:

1. Open the page you want to control
2. Click OpenClaw Relay icon in Chrome toolbar
3. Confirm badge is ON

4) One-pass convergence command pack

If you want a fast full sweep, run:

openclaw status
openclaw gateway status --deep
openclaw channels list
openclaw doctor --deep
openclaw gateway restart
openclaw logs --follow

Then send one Telegram message to your bot and click Relay once in Chrome. In most cases, logs will immediately reveal the exact failure point.

5) Stability recommendation: run OpenClaw on one 24/7 host

Many “random” outages are environment instability (laptop sleep, network switching, duplicate launches).

Production pattern:

• Run OpenClaw on exactly one always-on host
• Keep Telegram polling single-instance
• Use systemd with restart policy

Suggested deployment providers (2 vCPU / 2GB is enough to start):

• Tencent Cloud: strong connectivity for China users
• Vultr: broad global regions, flexible pricing
• DigitalOcean: straightforward docs, beginner-friendly ops

6) Bottom line

Most OpenClaw deployment failures are one of these:

• Conflict: multiple instances competing for token or port
• Connectivity: gateway/relay endpoint unreachable
• Configuration: token, webhook, pairing, bind/auth mismatch

Run this checklist end-to-end and you can usually restore service in 10–20 minutes.