tony/CHORUS
1
0
Fork 0
Code Issues 10 Pull Requests Actions Packages Projects Releases Wiki Activity

feat: Implement CHORUS scaling improvements for robust autoscaling

Browse Source 
Address WHOOSH issue #7 with comprehensive scaling optimizations to prevent
license server, bootstrap peer, and control plane collapse during fast scale-out.

HIGH-RISK FIXES (Must-Do):
✅ License gate already implemented with cache + circuit breaker + grace window
✅ mDNS disabled in container environments (CHORUS_MDNS_ENABLED=false)
✅ Connection rate limiting (5 dials/sec, 16 concurrent DHT queries)
✅ Connection manager with watermarks (32 low, 128 high)
✅ AutoNAT enabled for container networking

MEDIUM-RISK FIXES (Next Priority):
✅ Assignment merge layer with HTTP/file config + SIGHUP reload
✅ Runtime configuration system with WHOOSH assignment API support
✅ Election stability windows to prevent churn:
  - CHORUS_ELECTION_MIN_TERM=30s (minimum time between elections)
  - CHORUS_LEADER_MIN_TERM=45s (minimum time before challenging healthy leader)
✅ Bootstrap pool JSON support with priority sorting and join stagger

NEW FEATURES:
- Runtime config system with assignment overrides from WHOOSH
- SIGHUP reload handler for live configuration updates
- JSON bootstrap configuration with peer metadata (region, roles, priority)
- Configurable election stability windows with environment variables
- Multi-format bootstrap support: Assignment → JSON → CSV

FILES MODIFIED:
- pkg/config/assignment.go (NEW): Runtime assignment merge system
- docker/bootstrap.json (NEW): Example JSON bootstrap configuration
- pkg/election/election.go: Added stability windows and churn prevention
- internal/runtime/shared.go: Integrated assignment loading and conditional mDNS
- p2p/node.go: Added connection management and rate limiting
- pkg/config/hybrid_config.go: Added rate limiting configuration fields
- docker/docker-compose.yml: Updated environment variables and configs
- README.md: Updated status table with scaling milestone

This implementation enables wave-based autoscaling without system collapse,
addressing all scaling concerns from WHOOSH issue #7.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
anthonyrawlins
2025-09-23 17:50:40 +10:00
parent 26e4ef7d8b
commit e523c4b543
8 changed files with 776 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
p2p/node.go
View File

@@ -9,6 +9,7 @@ import (
"github.com/libp2p/go-libp2p"
"github.com/libp2p/go-libp2p/core/host"
"github.com/libp2p/go-libp2p/core/peer"
"github.com/libp2p/go-libp2p/p2p/net/connmgr"
"github.com/libp2p/go-libp2p/p2p/security/noise"
"github.com/libp2p/go-libp2p/p2p/transport/tcp"
kaddht "github.com/libp2p/go-libp2p-kad-dht"
@@ -44,13 +45,26 @@ func NewNode(ctx context.Context, opts ...Option) (*Node, error) {
listenAddrs = append(listenAddrs, ma)
}
// Create libp2p host with security and transport options
// Create connection manager with scaling-optimized limits
connManager, err := connmgr.NewConnManager(
config.LowWatermark, // Low watermark (32)
config.HighWatermark, // High watermark (128)
connmgr.WithGracePeriod(30*time.Second), // Grace period before pruning
)
if err != nil {
cancel()
return nil, fmt.Errorf("failed to create connection manager: %w", err)
}
// Create libp2p host with security, transport, and scaling options
h, err := libp2p.New(
libp2p.ListenAddrs(listenAddrs...),
libp2p.Security(noise.ID, noise.New),
libp2p.Transport(tcp.NewTCPTransport),
libp2p.DefaultMuxers,
libp2p.EnableRelay(),
libp2p.ConnectionManager(connManager), // Add connection management
libp2p.EnableAutoNATv2(), // Enable AutoNAT for container environments
)
if err != nil {
cancel()