openclaw - 💡(How to fix) Fix memory-core logs hourly "managed dreaming cron could not be reconciled" although cron service is running

Summary

After updating to openclaw 2026.5.9-beta.1, the gateway logs this warning every full hour:

memory-core: managed dreaming cron could not be reconciled (cron service unavailable).

The cron service itself appears to be running correctly. This looks like memory-core cannot resolve the cron service from the runtime/plugin hook context during its hourly reconciliation path, even though the gateway cron service is active.

Environment

• OpenClaw version: 2026.5.9-beta.1
• Commit: 0d3141ee24 fix(plugin-sdk): stabilize API baseline generation
• Container: openclaw-openclaw-gateway-1
• Host OS: Debian GNU/Linux 13 (trixie)
• Observed log timestamp: 2026-05-09T20:00:17.275+00:00

Evidence

Gateway log:

2026-05-09T20:00:17.275+00:00 [plugins] memory-core: managed dreaming cron could not be reconciled (cron service unavailable).

The cron service is available via CLI inside the gateway container:

docker exec openclaw-openclaw-gateway-1 openclaw cron list --json

Other cron jobs show normal state such as nextRunAtMs, lastRunStatus, and lastStatus.

The managed memory-core job exists and is healthy:

{
  "id": "6596fd0c-ac75-4624-8de8-181302bb5464",
  "name": "Memory Dreaming Promotion",
  "enabled": true,
  "schedule": { "kind": "cron", "expr": "0 3 * * *" },
  "sessionTarget": "isolated",
  "payload": {
    "kind": "agentTurn",
    "message": "__openclaw_memory_core_short_term_promotion_dream__",
    "lightContext": true
  },
  "state": {
    "nextRunAtMs": 1778382000000,
    "lastRunAtMs": 1778295685220,
    "lastRunStatus": "ok",
    "lastStatus": "ok",
    "lastDurationMs": 2049,
    "lastDeliveryStatus": "not-requested",
    "consecutiveErrors": 0,
    "consecutiveSkipped": 0
  },
  "status": "ok"
}

Expected behavior

If the gateway cron service is running and the managed Memory Dreaming Promotion job exists, memory-core should be able to reconcile the managed dreaming cron without logging cron service unavailable every hour.

Actual behavior

The warning appears every full hour, even though cron itself is active and the managed memory-core cron job exists with healthy state.

Initial diagnosis

This appears to be a regression in the plugin/runtime hook context rather than a local cron-store or scheduler failure.

Relevant observation from code inspection:

• memory-core uses ctx.getCron() during gateway_start and later runtime reconciliation.
• The managed job exists and the cron service can be queried from the gateway container.
• The warning seems to occur from the runtime reconciliation path when memory-core cannot resolve a CronServiceLike from the current hook context.

Possible area to inspect: whether before_agent_reply / runtime hook contexts should expose a live cron service, or whether memory-core should retain/use the gateway cron reference obtained during startup more reliably.

Impact

Severity: low-to-medium regression / noisy operational warning.

Cron jobs appear to continue running, but the hourly warning suggests memory-core cannot reliably reconcile its managed cron job after the update. This can make operational logs noisy and may hide a real managed-cron reconciliation failure if the job later needs updating.