Fix service status detection during container transitions

[Copilot]

_service_running() only checked for tasks in running state, causing false negatives during container startup/restart. This triggered premature demo shutdowns via the idle monitor when services were actually healthy but transitioning.

Changes

• _service_running(): Accept tasks in transitional states (new, pending, assigned, preparing, starting) in addition to running
• Edge case: Return True when replicas > 0 but no tasks exist yet (brief window during scale-up)
• Maintainability: Extract active states to ACTIVE_TASK_STATES constant

Before/After

# Before: Only checked running tasks
tasks = svc.tasks(filters={"desired-state": "running"})
return any(t.get("Status", {}).get("State") == "running" for t in tasks)

# After: Accept all non-terminal states
tasks = svc.tasks()
if not tasks and replicas > 0:
    return True  # Service scaling up
return any(t.get("Status", {}).get("State") in ACTIVE_TASK_STATES for t in tasks)

This prevents the heartbeat system from incorrectly detecting service failures during normal Docker Swarm lifecycle events.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>API doesnt read well demo's status</issue_title>
<issue_description>Sometimes when a demo is launched, ws stops, camera stops, as if the heartbeat went unsuccessful</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes API doesnt read well demo's status #3

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.