Summary

OpenClaw currently has no built-in monitoring for failed delivery queue items. When a message fails to send (e.g., Telegram message exceeds 4096 character limit), it causes:

1. Periodic gateway hangs - Failed messages retry every 15-30 minutes, blocking the gateway
2. No alerts - Users cannot detect queue jams proactively
3. Manual debugging required - Users must inspect logs and filesystem to identify issues

Steps to reproduce

1. Send a message longer than 4096 characters via Telegram (e.g., a detailed project report)
2. Observe the error in logs:

   [delivery-recovery] Recovery time budget exceeded — 1 entries deferred to next restart
   

3. Wait 15-30 minutes
4. Gateway becomes unresponsive due to retry attempts
5. Repeat indefinitely

Expected behavior

• ✅ Users should be notified when messages fail to deliver
• ✅ Failed items should have configurable retry limits
• ✅ Stuck messages should be auto-moved to dead-letter queue after N retries
• ✅ Gateway should provide queue status via CLI or API

Actual behavior

• ❌ No notification when messages fail
• ❌ Infinite retry loop (or until manual intervention)
• ❌ No CLI command to check queue status
• ❌ Gateway hangs periodically during retry attempts

OpenClaw version

2026.2.22-2

Operating system

macOS 26.3 (arm64)

Install method

Node: 25.6.1

Logs, screenshots, and evidence

Impact and severity

No response

Additional information

No response