Auto-provision TLS certificates during reconciliation

Replace the passive "log warning about missing certs" approach with active
auto-provisioning:

- When Cloudflare token and operator email are configured, automatically
  provision missing certs via certbot DNS-01 during reconciliation
- When credentials aren't available, log actionable guidance (no longer
  references a non-existent "Certificates page")
- Track TLS health in reconciler Health struct (ok/degraded with missing
  cert list)
- Broadcast tls:recovered SSE event when all certs become available
- Add CertManager interface and GetCloudflareToken callback to reconciler

The convergence loop (5 min) continuously retries failed provisions,
so transient DNS-01 failures self-heal on the next tick.

internal/api/v1/handlers.go

@@ -102,6 +102,8 @@ func NewAPI(dataDir, version string, allowedOrigins []string) (*API, error) {
 		SSE:                    sseManager,
 		StatePath:              filepath.Join(dataDir, "state.yaml"),
 		GenerateAutheliaConfig: api.generateAutheliaConfig,
+		Certs:                  api.certbot,
+		GetCloudflareToken:     api.getCloudflareToken,
 	}
 
 	// Wire up domain registration reconciliation: when domains change,

internal/reconcile/reconciler.go

@@ -50,6 +50,13 @@ type AuthManager interface {
 	RestartService() error
 }
 
+// CertManager is the subset of certbot.Manager used by reconciliation.
+type CertManager interface {
+	EnsureCredentials(apiToken string) error
+	Provision(domain, email string) error
+	BuildHAProxyCert(domain string) error
+}
+
 // DDNSRunner is the subset of ddns.Runner used by reconciliation.
 type DDNSRunner interface {
 	Trigger()
@@ -80,8 +87,10 @@ type SubsystemHealth struct {
 type Health struct {
 	HAProxy         SubsystemHealth `json:"haproxy"`
 	DNS             SubsystemHealth `json:"dns"`
+	TLS             SubsystemHealth `json:"tls"`
 	UpdatedAt       time.Time       `json:"updatedAt"`
 	ExcludedDomains []string        `json:"excludedDomains,omitempty"`
+	MissingCerts    []string        `json:"missingCerts,omitempty"`
 }
 
 // Reconciler orchestrates config regeneration when domains change.
@@ -92,11 +101,13 @@ type Reconciler struct {
 	HAProxy                HAProxyManager
 	DNS                    DNSManager
 	Auth                   AuthManager
+	Certs                  CertManager
 	DDNS                   DDNSRunner
 	DNSFilter              DNSFilterManager
 	SSE                    EventBroadcaster
 	StatePath              string
 	GenerateAutheliaConfig GenerateAutheliaConfigFn
+	GetCloudflareToken     func() string // callback to read CF token from secrets
 }
 
 // GetHealth returns a snapshot of the last reconciliation health state.
@@ -178,7 +189,19 @@ func (r *Reconciler) Reconcile() {
 	}
 
 	if len(httpRoutes) > 0 {
-		ensureTLSCerts(globalCfg, doms)
+		missingCerts := r.ensureTLSCerts(globalCfg, doms)
+		r.health.MissingCerts = missingCerts
+		if len(missingCerts) > 0 {
+			r.health.TLS = SubsystemHealth{
+				Status:  "degraded",
+				Message: fmt.Sprintf("%d certs missing: %v", len(missingCerts), missingCerts),
+			}
+		} else {
+			r.health.TLS = SubsystemHealth{Status: "ok"}
+		}
+	} else {
+		r.health.TLS = SubsystemHealth{Status: "ok"}
+		r.health.MissingCerts = nil
 	}
 
 	r.DDNS.Trigger()
@@ -192,6 +215,9 @@ func (r *Reconciler) Reconcile() {
 	if previousHealth.DNS.Status == "error" && r.health.DNS.Status == "ok" {
 		r.broadcastEvent("dnsmasq:recovered", "DNS recovered")
 	}
+	if previousHealth.TLS.Status != "ok" && r.health.TLS.Status == "ok" {
+		r.broadcastEvent("tls:recovered", "All TLS certificates provisioned")
+	}
 
 	slog.Info("networking updated", "component", "reconcile",
 		"domains", len(doms),
@@ -199,6 +225,7 @@ func (r *Reconciler) Reconcile() {
 		"l7Routes", len(httpRoutes),
 		"haproxy", r.health.HAProxy.Status,
 		"dns", r.health.DNS.Status,
+		"tls", r.health.TLS.Status,
 	)
 }
 
@@ -395,35 +422,83 @@ func (r *Reconciler) broadcastDNSEvent(eventType, message string) {
 	})
 }
 
-// ensureTLSCerts logs which certificates are missing for registered domains.
-// Does NOT auto-provision — cert provisioning is user-initiated.
-func ensureTLSCerts(globalCfg *config.State, doms []domains.Domain) {
+// ensureTLSCerts checks for missing certificates and auto-provisions them
+// when Cloudflare credentials and operator email are available.
+// Returns the list of domains still missing certs after provisioning attempts.
+func (r *Reconciler) ensureTLSCerts(globalCfg *config.State, doms []domains.Domain) []string {
 	gatewayDomain := ""
 	if parts := strings.SplitN(globalCfg.Cloud.Central.Domain, ".", 2); len(parts) == 2 {
 		gatewayDomain = parts[1]
 	}
 
+	// Collect unique missing certs (wildcard or individual)
+	missing := map[string]bool{}
 	for _, dom := range doms {
 		if dom.TLS != domains.TLSTerminate || dom.DomainName == "" {
 			continue
 		}
-
-		// Check if covered by wildcard
 		if gatewayDomain != "" && strings.HasSuffix(dom.DomainName, "."+gatewayDomain) {
-			wildcardCert := certbot.HAProxyCertPath(gatewayDomain)
-			if _, err := os.Stat(wildcardCert); err != nil {
-				slog.Warn("no wildcard cert for gateway domain — provision via Certificates page", "component", "reconcile",
-					"domain", dom.DomainName, "needed", "*."+gatewayDomain)
+			if _, err := os.Stat(certbot.HAProxyCertPath(gatewayDomain)); err != nil {
+				missing["*."+gatewayDomain] = true
 			}
 			continue
 		}
-
-		// Check individual cert
 		if _, err := os.Stat(certbot.HAProxyCertPath(dom.DomainName)); err != nil {
-			slog.Warn("no cert for domain — provision via Certificates page", "component", "reconcile",
-				"domain", dom.DomainName)
+			missing[dom.DomainName] = true
 		}
 	}
+
+	if len(missing) == 0 {
+		return nil
+	}
+
+	// Can we auto-provision?
+	email := globalCfg.Operator.Email
+	cfToken := ""
+	if r.GetCloudflareToken != nil {
+		cfToken = r.GetCloudflareToken()
+	}
+
+	if cfToken == "" || email == "" || r.Certs == nil {
+		// Can't auto-provision — report what's missing
+		var names []string
+		for domain := range missing {
+			names = append(names, domain)
+			slog.Warn("missing cert (auto-provision unavailable — configure Cloudflare token and operator email)",
+				"component", "reconcile", "domain", domain)
+		}
+		return names
+	}
+
+	// Auto-provision missing certs
+	if err := r.Certs.EnsureCredentials(cfToken); err != nil {
+		slog.Error("failed to write certbot credentials", "component", "reconcile", "error", err)
+		var names []string
+		for domain := range missing {
+			names = append(names, domain)
+		}
+		return names
+	}
+
+	var stillMissing []string
+	for domain := range missing {
+		provisionDomain := domain
+		if strings.HasPrefix(domain, "*.") {
+			provisionDomain = domain // certbot handles wildcard with -d *.example.com
+		}
+
+		slog.Info("auto-provisioning certificate", "component", "reconcile", "domain", provisionDomain)
+		if err := r.Certs.Provision(provisionDomain, email); err != nil {
+			slog.Error("cert auto-provision failed", "component", "reconcile", "domain", provisionDomain, "error", err)
+			stillMissing = append(stillMissing, domain)
+		} else {
+			slog.Info("certificate provisioned", "component", "reconcile", "domain", provisionDomain)
+			// Build HAProxy PEM for non-wildcard certs
+			baseDomain := strings.TrimPrefix(domain, "*.")
+			_ = r.Certs.BuildHAProxyCert(baseDomain)
+		}
+	}
+	return stillMissing
 }
 
 // hasCertForDomain checks if a valid (non-empty) cert exists for a domain —