# 2026-05-01 — Daily Log## CPU Drain Investigation & Fix- OpenClaw gateway was at 61% CPU at idle after v2026.4.29 upgrade attempt- Root cause: Discord crash loop — rate-limited slash commands, gateway timeouts, 30-40s event loop blocks- Discord was reconnecting in a loop, each cycle blocking Node event loop- Agent bootstraps taking 80-117s (should be 5-10s), LLM calls timing out, Telegram send failures- Fix: Disabled Discord in openclaw.json (`channels.discord.enabled = false`)- CPU dropped from 61% → 14% after Discord disable + gateway restart- Also set `channels.telegram.timeoutSeconds = 30` (v2026.4.8 didn't have v2026.4.29's polling fix)## OpenClaw Upgrade Status- Claw attempted upgrade to v2026.4.29 — went sideways- npm still shows v2026.4.8 installed globally- Partial artifacts remain: plugin-runtime-deps/openclaw-2026.4.29-*, openclaw.json.backup-2026.4.29, openclaw.json.last-good- Config lastTouchedVersion = 2026.4.29 but actual installed version = 2026.4.8- Need to decide: complete upgrade or stay on v2026.4.8## New Software Found- Claw Empire (climpire v2.0.4) installed at ~/openclaw/claw-empire/ — multi-agent orchestration dashboard- Not currently running, no systemd service- Mission Control (robsannaa) also not running — was on port 3333 dev mode## Pi Reboot- Topher rebooted the Pi during the session- Pi back online on Tailscale (100.99.44.24, 11ms)