Core Changes:
- Add Agent struct and management methods to Store interface
- Implement PostgreSQL store methods (UpsertAgent, ListAgents, DeleteAgent, DeleteStaleAgents)
- Integrate persistence into Registry with async database writes
- Add Load() method to restore agents from database on startup
- Implement runAgentCleanup background task (5min interval, 10min stale threshold)
Database:
- Update agents table schema to use JSONB for groups field
- Add indexes on last_heartbeat and healthy columns
- Support health tracking and automatic cleanup of stale agents
Documentation:
- Add comprehensive DB access and upgrade guide
- Include agent persistence implementation plan
- Document diagnostic procedures and troubleshooting steps
- Add walkthrough of multi-agent support implementation
This enables:
- Persistent agent state across service restarts
- Automatic cleanup of offline agents
- Multi-agent support with shared token authentication
- Add AgentID field to StatusReport for agent self-identification
- Implement RegisterAgent() method for dynamic agent registration
- Modify agentReportStatusHandler to extract and use agent-reported IDs
- Use wildcard credential to support multiple agents with INTERNAL_SERVICE_TOKEN
- Add agents table schema for future persistence
This enables scalable multi-node deployment without pre-configuring each agent ID.
