Telegram → Agents Hub message sync not responding due to API rate limits
What We Built:
- Cross-platform message sync protocol documented in AGENTS.md
- Backend integration module: backend/src/integrations/telegram-sync.js
- Protocol: POST user messages (channel=dashboard) and agent responses (channel=direct) to
/api/os/messages
- Tests pass:
message sync task creation status updates
Current Problem:
message-inbox cron hitting Anthropic API rate limits every minute:
Error: Provider anthropic is in cooldown (rate_limit)
Root Cause:
4 crons running too frequently on same API account:
- message-inbox: every 1min
- task-watcher: every 3min
- team-heartbeat: every 5min
- vercel-job: every 2min
Result: 5 unread messages stuck in database, cron can't respond.
Solutions:
Option 1 (Quick): Reduce message-inbox to every 5min instead of 1min
Option 2 (Better): Stagger cron schedules:
message-inbox: */5 (0,5,10...)
task-watcher: 1-59/5 (1,6,11...)
team-heartbeat: 2-59/5 (2,7,12...)
vercel-job: 3-59/5 (3,8,13...)
Option 3 (Best): Add exponential backoff on rate limit errors (30s, 60s, skip)
Action Required:
Implement Option 1 immediately to restore service, then Option 2 for reliability.
Files Modified:
- AGENTS.md - protocol docs
- backend/src/integrations/telegram-sync.js - working sync module
- backend/src/integrations/TELEGRAM_INTEGRATION.md - implementation guide
- Cross-platform message sync protocol documented in AGENTS.md
- Backend integration module: backend/src/integrations/telegram-sync.js
- Protocol: POST user messages (channel=dashboard) and agent responses (channel=direct) to
/api/os/messages
- Tests pass:
message sync task creation status updatesCurrent Problem:
message-inbox cron hitting Anthropic API rate limits every minute:
Error: Provider anthropic is in cooldown (rate_limit)
Root Cause:
4 crons running too frequently on same API account:
- message-inbox: every 1min
- task-watcher: every 3min
- team-heartbeat: every 5min
- vercel-job: every 2min
Result: 5 unread messages stuck in database, cron can't respond.
Solutions:
Option 1 (Quick): Reduce message-inbox to every 5min instead of 1min
Option 2 (Better): Stagger cron schedules:
message-inbox: */5 (0,5,10...)
task-watcher: 1-59/5 (1,6,11...)
team-heartbeat: 2-59/5 (2,7,12...)
vercel-job: 3-59/5 (3,8,13...)
Option 3 (Best): Add exponential backoff on rate limit errors (30s, 60s, skip)
Action Required:
Implement Option 1 immediately to restore service, then Option 2 for reliability.
Files Modified:
- AGENTS.md - protocol docs
- backend/src/integrations/telegram-sync.js - working sync module
- backend/src/integrations/TELEGRAM_INTEGRATION.md - implementation guide
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
83,815Members
Resources
Similar Threads
Was this page helpful?
