Operator-visible logging for the live runner lifecycle

What

Add operator-visible logging for the live runner lifecycle, so "is my runner connected, and at what price?" can be read from logs instead of inferred from /discovery.

Lifecycle events to surface:

Dynamic runners

• live runner registered — new-runner branch of Heartbeat
• live runner deregistered — with a reason (heartbeat timeout on crash/OOM/network loss/hard kill, unregister on graceful SDK shutdown)

Static runners

• static live runner registered — first registration only
• static live runner health changed — on ready <-> unavailable transitions

Pricing

• live runner price set — debug level, on first set and on actual change

Why

Better logging makes the runner lifecycle observable for developers and operators running the live runner via the Python SDK.

Implementation

The go-livepeer side is implemented in livepeer/go-livepeer#3943.