feat: Docker health check for bot container readiness

[interplanetarychris]

Summary

Add Docker HEALTHCHECK to bot containers so the dashboard can show accurate startup status instead of using a time-based heuristic.

Problem

Previously, the dashboard used an 8-second timer to guess when a bot was ready. This was unreliable - OpenClaw can take 10+ seconds to start, leading to "running" status while the bot was still initializing.

Solution

• Add Docker HEALTHCHECK config when creating bot containers (polls /health endpoint)
• Extract health status from Docker inspect API
• Dashboard uses actual health status (starting → healthy) instead of time heuristic
• Containers without healthcheck (created before this update) show as "running" immediately

Changes

File	Change
src/services/DockerService.ts	Add HEALTHCHECK config, extract health from inspect
src/types/container.ts	Add HealthStatus type and health field
dashboard/src/types.ts	Add HealthStatus type and health field
dashboard/src/dashboard/BotCard.tsx	Use health status instead of 8s timer

Health check config

Healthcheck: {
  Test: ['CMD', 'wget', '-q', '--spider', 'http://localhost:{port}/health'],
  Interval: 2_000_000_000,  // 2s
  Timeout: 3_000_000_000,   // 3s
  Retries: 30,
  StartPeriod: 5_000_000_000,  // 5s
}

Test plan

• Create new bot - verify "starting" shows until /health responds
• Existing bots (no healthcheck) show "running" immediately
• Unhealthy containers show error status

🤖 Generated with Claude Code