fix: prevent WS-broadcast OOM crash under connection churn

[skialpine]

Stacked on feat/scale-telemetry (PR decentespresso#57). Base will be retargeted to main once decentespresso#57 merges. The telemetry in decentespresso#57 (reset_reason, heap) is what made this diagnosable.

Root cause (from a captured + decoded panic backtrace)

Under sustained multi-client WiFi load (WS connection churn + the 10 Hz weight broadcast), free heap collapses. The broadcast path then allocates an AsyncWebSocketMessage per client and operator new throws std::bad_alloc; Arduino-ESP32 builds -fno-exceptions, so the throw goes to std::terminate() → abort() → reboot (reset_reason=panic). This is the "weight stops being collected under load" failure — not thermal (die temp was 33 °C).

Decoded stack:

operator new -> __cxa_throw -> std::terminate -> abort
AsyncWebSocketClient::_queueMessage   (AsyncWebSocket.cpp:490)
AsyncWebSocket::printfAll
sendWebsocketWeightAll                 (include/websocket.h, loop() 10 Hz broadcast)

The existing 15 KB heap watchdog can't prevent it: it has a 2 s debounce and defers reboot up to 60 s while BLE is connected, so the 10 Hz allocation bad_allocs long before it acts.

Fix

• wsBroadcastHeapOk() heap-floor gate on every broadcast-to-all helper (sendWebsocketWeightAll, sendWebsocketStatusAll, button, power-off): when free heap is below WS_BROADCAST_HEAP_FLOOR (25 KB, above the 15 KB watchdog) the frame is skipped, not allocated. Dropping a frame is invisible (next weight frame ≤500 ms away).
• -D WS_MAX_QUEUED_MESSAGES=8 (lib default 32): bounds each client's outbound queue so a backed-up/half-open client can't hoard heap.
• CLAUDE.md: documented the footgun (notes + troubleshooting table).

Verification (on hardware)

Re-ran the exact load that crashed the unpatched build — conn_churn --rst 8×8 + 10 Hz WS + mDNS, BT connected:

• Free heap driven to 6436 bytes (old build crashed at ~4684).
• Gate engaged: [ws] low heap 17736 < 25000 -> skip broadcast.
• No abort, no reboot, weight stream uninterrupted (uptime continuous).

🤖 Generated with Claude Code