๐ Bug Report: Gateway crashes on transient network failures
Bug Report: Gateway crashes on transient network failuresVersion: Clawdbot 2026.1.24-3
Node: v25.4.0
OS: macOS 15.6 (arm64)
Setup: Mac Mini running 24/7 as always-on gateway via launchd
Problem:
Gateway process crashes when any fetch() call fails due to transient network issues. Instead of catching the error gracefully, it throws an unhandled promise rejection and the process exits.
Error from logs:
[clawdbot] Unhandled promise rejection: TypeError: fetch failed
at node:internal/deps/undici/undici:16416:13
at processTicksAndRejections (node:internal/process/task_queues:104:5)
Frequency: 10+ crashes today alone. Timestamps:
โข 15:36, 16:06, 16:29, 17:52, 18:46, 18:53, 19:51, 20:21, 20:45, 20:57 UTC
Impact:
โข Gateway restarts every time (launchd auto-restart)
โข Session context is lost on each restart
โข Boot hooks fire repeatedly (spam notifications)
โข User experience degraded
Root cause:
Network blips (even sub-second) cause fetch() to fail. The fetch call somewhere in the codebase doesn't have try/catch or .catch() handler.
Expected behavior:
Transient network errors should be caught and retried, or at minimum logged without crashing the process. A fetch failure shouldn't bring down the entire gateway.
Suggested fix:
1. Wrap all fetch() calls in try/catch
2. Add global unhandledRejection handler that logs but doesn't exit
3. Implement retry logic for API calls
Happy to provide more logs if needed!
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
85,605Members
Resources
Similar Threads
Was this page helpful?
