Description

When the gateway process becomes unresponsive (possibly due to network timeouts or hanging async operations), it holds onto port 18789 but stops responding to systemd's stop signals. This causes an infinite restart loop where:

1. Old process holds port 18789 (zombie state)
2. Systemd tries to start new instance
3. New instance sees port in use → exits with code 1
4. Systemd waits 5s → restart
5. Repeat indefinitely (observed 3400+ restarts over ~24h)

Environment

• Moltbot version: latest (installed via git)
• Node: v22.22.0
• OS: Ubuntu 24.04 (Linux 6.8.0-90-generic)
• Systemd service: Type=simple with Restart=always

Logs

Before crash-loop, logs show repeated:

[moltbot] Suppressed AbortError: AbortError: This operation was aborted
    at node:internal/deps/undici/undici:14902:13
[telegram] network error: Request to 'getUpdates' timed out after 30 seconds
[tools] cron failed: gateway timeout after 10000ms

Then crash-loop:

Gateway failed to start: gateway already running (pid XXXXX); lock timeout after 5000ms
Port 18789 is already in use.

Workaround

Added to systemd service:

TimeoutStopSec=30
KillMode=mixed
SendSIGKILL=yes
StartLimitIntervalSec=300
StartLimitBurst=10

Suggested Fix

1. Gateway should handle SIGTERM gracefully and release port quickly
2. Consider adding a health check endpoint
3. Maybe use SO_REUSEADDR or implement proper graceful shutdown

Impact

• Messages truncated mid-response ("terminated" sent to users)
• Gateway unresponsive for extended periods
• Log files grow very large (2.9MB+ of repeated error messages)