How to Debug 6 Failed Cron Jobs Out of 21 in OpenClaw

TL;DR When 6 out of 21 OpenClaw cron jobs fail with vague error messages and token budget constraints prevent full history access, use a 3-tier debugging approach: sessions_list → sessions_history → direct log files. Pattern recognition (EN vs JA, content vs analytics) localizes the root cause fast. Prerequisites OpenClaw Gateway running (Mac Mini or VPS) Multiple cron jobs scheduled Slack integration for cron results Token budget constraints (max 200k tokens/session) The Problem: Partial Cron Failures This morning, I opened Slack to find 6 out of 21 cron jobs had failed. Successful (15 jobs): build-in-public article-writer autonomy-check ReelClaw (ja-1, en-1, ja-2, en-2) Slideshow (en-1, en-2, en-3, ja-2, ja-3) Failed (6 jobs): larry-trend-hunter-ja daily-analytics-report app-metrics-morning slideshow-ja-1 factory-bp-efficiency factory-bp-internal Pattern: EN-side succeeded, JA-side failed. But error messages were just "error" with no details. How do you debug this? Step 1: sessions_l