chore(agent): log agent-server boot timing (bootMs)

[tatoalo]

Problem

Cloud-task agent-server spawn time, from node process start to a usable session (hasSession=true, what the backend health poll waits on) is currently unmeasured, so we're optimizing it blind. We want to land lightweight timing first and verify where boot time actually goes in production before shipping any speedups.

Changes

Add bootMs (wall-clock since process start, via process.uptime()) to two existing agent-server log lines:

• "HTTP server listening" — process start → HTTP accepting (node boot + bundle/SDK parse)
• "Session initialized successfully" — process start → hasSession=true (the full spawn)

[github-actions]

React Doctor found no issues in the changed files. 🎉

_{Reviewed by React Doctor for commit ebae605.}

[greptile-apps]

_{Reviews (1): Last reviewed commit: "chore(agent): log agent-server boot timi..." | Re-trigger Greptile}

[greptile-apps]

_{Reviews (2): Last reviewed commit: "chore(agent): update /health test for bo..." | Re-trigger Greptile}