fix: recover faster after network outage (fixes #49)

[Hany-Almnaem]

This PR improves SPARK's ability to recover quickly after network outages.

Changes:

• Introduces ActivityState.isHealthy() to expose health state
• Modifies spark.js#run() to:
  • Retry every 5 seconds when offline (OFFLINE_RETRY_DELAY_MS)
  • Use jittered delay when healthy (via calculateDelayBeforeNextTask)
• Adds a test case to verify offline retry delay

This makes the station feel more responsive after waking from sleep or reconnecting.

All tests pass
Fixes #49

[bajtos]

Thank you for the pull request, @Hany-Almnaem! This will be a great and impactful improvement. ❤️

[bajtos]

Almost there!