feat: Add WebSocket support for real-time TUI updates

[stormcat24]

Summary

Add WebSocket support to enable real-time updates in the TUI, eliminating delays when resources are created, updated, or deleted.

Current Problem

Currently, the TUI polls for updates every 5 seconds, which causes noticeable delays:

• When a cluster is created, it takes up to 10-30 seconds to appear in the TUI
• Task status changes are not reflected immediately
• Service deployments don't show up right away

This creates a poor user experience where users have to wait or manually refresh to see their changes.

Root Causes of Delays

1. TUI polling interval: 5 seconds (in internal/tui/model.go)
2. Batch update delay: 2 seconds (sync controller's BatchUpdater)
3. Initial pod processing delay: 3 seconds (processExistingPods sleep)

Proposed Solution

Implement WebSocket-based real-time notifications:

Architecture

1. WebSocket Server

   • Add WebSocket endpoint to the admin server (port 5374)
   • Path: /ws/events or similar
   • Broadcast events for resource changes (clusters, services, tasks, etc.)

2. Event Types

   type Event struct {
       Type      string    // "cluster.created", "task.updated", etc.
       Resource  string    // Resource ARN
       Data      interface{} // Resource data
       Timestamp time.Time
   }

3. TUI WebSocket Client

   • Connect to WebSocket on startup
   • Handle reconnection logic
   • Update local state immediately when events are received
   • Fall back to polling if WebSocket is unavailable

Implementation Areas

• Add WebSocket handler to admin server (internal/controlplane/admin/)
• Create event broadcaster/hub for managing connections
• Emit events from API handlers (CreateCluster, CreateService, etc.)
• Add WebSocket client to TUI (internal/tui/)
• Handle WebSocket messages in TUI update loop
• Implement reconnection logic with exponential backoff
• Add feature flag to enable/disable WebSocket mode

Benefits

• Immediate feedback: Resources appear instantly after creation
• Reduced server load: Less frequent polling required
• Better UX: Real-time updates for all resource changes
• Foundation for future features: Live logs, metrics, etc.

Alternative Short-term Solutions

While this is being implemented, we can:

1. Reduce TUI polling interval from 5s to 2s
2. Add immediate refresh after create/update operations in TUI
3. Add manual refresh command (e.g., Ctrl+R)

Related Issues

• Could also benefit from GraphQL subscriptions as an alternative
• Would enable real-time log streaming in the future

Technical Considerations

• Use gorilla/websocket (already in dependencies)
• Consider using a message queue for reliability
• Implement proper connection pooling and cleanup
• Add metrics for WebSocket connections

Acceptance Criteria

• TUI connects to WebSocket on startup
• Resources appear in TUI within 100ms of creation
• Graceful fallback to polling if WebSocket fails
• No memory leaks from connection management
• Works across all supported platforms

Priority

Medium - This is a quality of life improvement that would significantly enhance the user experience, but the current polling mechanism is functional.