Description

Gateway crashes when a network request fails, instead of handling the error gracefully.

Error Log

19:46:07 [clawdbot] Non-fatal rejection (continuing): fetch failed
19:46:07 [clawdbot] Unhandled promise rejection: TypeError: fetch failed at node:internal/deps/undici/undici:13510:13

Context

• The gateway was running normally (tick and health events working fine)
• At 19:46:07, a fetch request failed (possibly to an LLM API, Telegram, or cron job like BibiGPT)
• The promise rejection was not caught, causing the gateway to crash

Expected Behavior

Network failures should be caught and handled gracefully. The gateway should:

1. Log the error
2. Retry if appropriate
3. Continue running (not crash)

Environment

• OS: Windows Server 2019
• Node: v22.18.0
• Clawdbot version: 2026.1.24-3

Additional Context

This happened in the morning when opening the gateway. There was also a separate tool_use_id not found error earlier, but this fetch failure appears to be a different issue - pure network error handling.