Scheduled tasks / cron jobs don't fire in Cloudchamber containers

[MichaelMichaelMichaelMichaelMichael]

Setup

Running OpenClaw (AI assistant) in a Moltworker-managed Cloudchamber at Cloudflare (worker setup) container with:

• Persistent storage via /data (R2)
• Telegram webhook for incoming messages
Problem

OpenClaw has a built-in cron scheduler that runs inside the Gateway process. Jobs are stored correctly and the scheduler reports enabled: true, but jobs never fire automatically.

Observations:

• Gateway process runs for hours, zero automatic cron executions
• Manual trigger via CLI (openclaw cron run --force) works instantly
• Incoming Telegram messages wake the container and are processed fine
• After extended inactivity, the container stops responding entirely and seems to lose its Moltworker identity/registration (unless that's a separate issue)
Suspected cause

Cloudchamber appears to suspend containers during idle periods, freezing the Node.js event loop. Internal timers (setTimeout/setInterval) don't tick while suspended. The container only wakes on incoming HTTP requests.

Question

Is there a recommended pattern for scheduled/background tasks in Cloudchamber?

Options I'm considering:

1. External Cloudflare Cron Trigger to ping the container periodically
2. Moltworker configuration for keep-alive / minimum instances
3. Different architecture approach
   Any guidance appreciated!