perf: enhance monitor stability with non-blocking LOCK acquisition (#1754)

[zhu6201976]

Summary:
While the recent changes in develop introduce a global wait_for timeout for update_timeline, the monitor still attempts to "queue" for the global LOCK. Under extreme load, this can lead to unnecessary resource contention.

Improvements:

Non-blocking Stats: Changed LOCK acquisition within monitor.py to a fail-fast mechanism (0.5s timeout). If the crawler pool is busy, the monitor skips the current data point instead of adding to the lock queue.

System Resilience: Ensures that telemetry collection never interferes with the core mission-critical path (browser lifecycle management).

Error Silencing: Downgraded lock-timeout warnings to debug logs to avoid cluttering production logs during expected high-load spikes.

Why this matters:
In high-concurrency enterprise environments, skipping a single monitoring heartbeat is preferable to adding latency to the asyncio event loop.