[Monitoring] Define and implement health check endpoints for deployed instances

[thealphacubicle]

Overview

Define and implement a health check endpoint for OpenContext MCP server deployments.
This is important for the dashboard, the Massachusetts MCP Registry, and general
ops visibility — the registry needs something to poll to verify an instance is live.

Goals

• Define what a healthy instance looks like (Lambda reachable, plugin loaded,
  upstream data portal reachable)
• Implement a /health or MCP-compatible ping endpoint on the Lambda handler
• Ensure the endpoint returns meaningful status (not just HTTP 200 — include
  plugin name, version, upstream portal status)
• Document the health check contract so the registry and dashboard can consume it

Deliverables

• Health check response schema defined and documented
• Health check endpoint implemented in Lambda handler
• Upstream portal reachability check included in response
• Unit test covering health check response shape
• docs/DEPLOYMENT.md updated with health check endpoint details

Notes

• Should be unauthenticated — health checks need to work before OAuth is in place