mcp/internal/config/agent.go

package config

import (
	"fmt"
	"os"
	"time"

	toml "github.com/pelletier/go-toml/v2"
)

// AgentConfig is the configuration for the mcp-agent daemon.
type AgentConfig struct {
	Server   ServerConfig   `toml:"server"`
	Database DatabaseConfig `toml:"database"`
	MCIAS    MCIASConfig    `toml:"mcias"`
	Agent    AgentSettings  `toml:"agent"`
	MCProxy  MCProxyConfig  `toml:"mcproxy"`
	Monitor  MonitorConfig  `toml:"monitor"`
	Log      LogConfig      `toml:"log"`
}

// MCProxyConfig holds the mc-proxy connection settings.
type MCProxyConfig struct {
	// Socket is the path to the mc-proxy gRPC admin API Unix socket.
	// If empty, route registration is disabled.
	Socket string `toml:"socket"`

	// CertDir is the directory containing TLS certificates for routes.
	// Convention: <service>.pem and <service>.key per service.
	// Defaults to /srv/mc-proxy/certs.
	CertDir string `toml:"cert_dir"`
}

// ServerConfig holds gRPC server listen address and TLS paths.
type ServerConfig struct {
	GRPCAddr string `toml:"grpc_addr"`
	TLSCert  string `toml:"tls_cert"`
	TLSKey   string `toml:"tls_key"`
}

// DatabaseConfig holds the SQLite database path.
type DatabaseConfig struct {
	Path string `toml:"path"`
}

// AgentSettings holds agent-specific settings.
type AgentSettings struct {
	NodeName         string `toml:"node_name"`
	ContainerRuntime string `toml:"container_runtime"`
}

// MonitorConfig holds monitoring loop parameters.
type MonitorConfig struct {
	Interval      Duration `toml:"interval"`
	AlertCommand  []string `toml:"alert_command"`
	Cooldown      Duration `toml:"cooldown"`
	FlapThreshold int      `toml:"flap_threshold"`
	FlapWindow    Duration `toml:"flap_window"`
	Retention     Duration `toml:"retention"`
}

// LogConfig holds logging settings.
type LogConfig struct {
	Level string `toml:"level"`
}

// Duration wraps time.Duration to support TOML string unmarshaling.
// It accepts Go duration strings (e.g. "60s", "15m", "24h") plus a
// "d" suffix for days (e.g. "30d" becomes 30*24h).
type Duration struct {
	time.Duration
}

// UnmarshalText implements encoding.TextUnmarshaler for TOML parsing.
func (d *Duration) UnmarshalText(text []byte) error {
	s := string(text)
	if s == "" {
		d.Duration = 0
		return nil
	}

	// Handle "d" suffix for days.
	if len(s) > 1 && s[len(s)-1] == 'd' {
		var days float64
		if _, err := fmt.Sscanf(s[:len(s)-1], "%f", &days); err != nil {
			return fmt.Errorf("parse duration %q: %w", s, err)
		}
		d.Duration = time.Duration(days * float64(24*time.Hour))
		return nil
	}

	dur, err := time.ParseDuration(s)
	if err != nil {
		return fmt.Errorf("parse duration %q: %w", s, err)
	}
	d.Duration = dur
	return nil
}

// MarshalText implements encoding.TextMarshaler for TOML serialization.
func (d Duration) MarshalText() ([]byte, error) {
	return []byte(d.String()), nil
}

// LoadAgentConfig reads and validates an agent configuration file.
// Environment variables override file values for select fields.
func LoadAgentConfig(path string) (*AgentConfig, error) {
	data, err := os.ReadFile(path) //nolint:gosec // config path from trusted CLI flag
	if err != nil {
		return nil, fmt.Errorf("read config %q: %w", path, err)
	}

	var cfg AgentConfig
	if err := toml.Unmarshal(data, &cfg); err != nil {
		return nil, fmt.Errorf("parse config %q: %w", path, err)
	}

	applyAgentDefaults(&cfg)
	applyAgentEnvOverrides(&cfg)

	if err := validateAgentConfig(&cfg); err != nil {
		return nil, fmt.Errorf("validate config: %w", err)
	}

	return &cfg, nil
}

func applyAgentDefaults(cfg *AgentConfig) {
	if cfg.Monitor.Interval.Duration == 0 {
		cfg.Monitor.Interval.Duration = 60 * time.Second
	}
	if cfg.Monitor.Cooldown.Duration == 0 {
		cfg.Monitor.Cooldown.Duration = 15 * time.Minute
	}
	if cfg.Monitor.FlapThreshold == 0 {
		cfg.Monitor.FlapThreshold = 3
	}
	if cfg.Monitor.FlapWindow.Duration == 0 {
		cfg.Monitor.FlapWindow.Duration = 10 * time.Minute
	}
	if cfg.Monitor.Retention.Duration == 0 {
		cfg.Monitor.Retention.Duration = 30 * 24 * time.Hour // 30 days
	}
	if cfg.Log.Level == "" {
		cfg.Log.Level = "info"
	}
	if cfg.Agent.ContainerRuntime == "" {
		cfg.Agent.ContainerRuntime = "podman"
	}
	if cfg.MCProxy.CertDir == "" {
		cfg.MCProxy.CertDir = "/srv/mc-proxy/certs"
	}
}

func applyAgentEnvOverrides(cfg *AgentConfig) {
	if v := os.Getenv("MCP_AGENT_SERVER_GRPC_ADDR"); v != "" {
		cfg.Server.GRPCAddr = v
	}
	if v := os.Getenv("MCP_AGENT_SERVER_TLS_CERT"); v != "" {
		cfg.Server.TLSCert = v
	}
	if v := os.Getenv("MCP_AGENT_SERVER_TLS_KEY"); v != "" {
		cfg.Server.TLSKey = v
	}
	if v := os.Getenv("MCP_AGENT_DATABASE_PATH"); v != "" {
		cfg.Database.Path = v
	}
	if v := os.Getenv("MCP_AGENT_NODE_NAME"); v != "" {
		cfg.Agent.NodeName = v
	}
	if v := os.Getenv("MCP_AGENT_CONTAINER_RUNTIME"); v != "" {
		cfg.Agent.ContainerRuntime = v
	}
	if v := os.Getenv("MCP_AGENT_LOG_LEVEL"); v != "" {
		cfg.Log.Level = v
	}
	if v := os.Getenv("MCP_AGENT_MCPROXY_SOCKET"); v != "" {
		cfg.MCProxy.Socket = v
	}
	if v := os.Getenv("MCP_AGENT_MCPROXY_CERT_DIR"); v != "" {
		cfg.MCProxy.CertDir = v
	}
}

func validateAgentConfig(cfg *AgentConfig) error {
	if cfg.Server.GRPCAddr == "" {
		return fmt.Errorf("server.grpc_addr is required")
	}
	if cfg.Server.TLSCert == "" {
		return fmt.Errorf("server.tls_cert is required")
	}
	if cfg.Server.TLSKey == "" {
		return fmt.Errorf("server.tls_key is required")
	}
	if cfg.Database.Path == "" {
		return fmt.Errorf("database.path is required")
	}
	if cfg.MCIAS.ServerURL == "" {
		return fmt.Errorf("mcias.server_url is required")
	}
	if cfg.MCIAS.ServiceName == "" {
		return fmt.Errorf("mcias.service_name is required")
	}
	if cfg.Agent.NodeName == "" {
		return fmt.Errorf("agent.node_name is required")
	}
	return nil
}