feat(eda): add webhook, polling, and Kafka event source plugins

[stevefulme1]

Summary

Adds Event-Driven Ansible (EDA) source plugins to enable bidirectional automation with Edwin AI. Edwin AI can now trigger Ansible playbooks in response to alerts, complementing the existing data collection modules.

Event Source Plugins

Plugin	Pattern	Use Case
webhook	Edwin AI pushes to EDA	Real-time, simplest setup
alerts	EDA polls Edwin AI API	Works behind firewalls, no Edwin config needed
kafka	EDA consumes from Kafka	High-volume enterprise deployments

All three sources emit a consistent normalized event schema under the edwin_ai key:

edwin_ai:
  event_type: alert
  alert_id: "12345"
  severity: critical
  host: web-server-01
  message: CPU usage exceeded 95%
  metric: cpu.usage
  value: 97.3
  threshold: 95
  resource_group: production
  timestamp: 1715788800
  status: active
  raw: {original payload}

Example Rulebooks

• alert_remediation.yml — webhook-based auto-remediation (CPU, disk, memory, service down)
• alert_polling.yml — polling-based with incident creation/closure
• alert_kafka.yml — Kafka stream processing with escalation

Auth

Uses the same portal/access_id/access_key pattern as the existing query_api module. Webhook source optionally supports HMAC signature verification.

Dependencies

• webhook: aiohttp (included in EDA controller)
• alerts: requests (already a collection dependency)
• kafka: aiokafka (optional, only needed for Kafka source)

Test plan

• Webhook source receives and normalizes sample payloads
• Alerts source authenticates and polls Edwin AI API
• Kafka source connects and consumes from topic
• Rulebook conditions match on normalized event fields
• HMAC signature verification rejects invalid webhooks