fix: initialize counter and histogram metrics at startup

Counter and histogram metrics were absent from Prometheus scrapes until
the first deployment occurred, making it impossible to distinguish
"no deployments" from "exporter not running" in dashboards and alerts.

Initialize all expected label combinations with zero values when the
collector is created so metrics appear in every scrape from startup.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

internal/metrics/metrics.go

@@ -53,9 +53,36 @@ func NewCollector(reg prometheus.Registerer) *Collector {
 	reg.MustRegister(c.deploymentInProgress)
 	reg.MustRegister(c.info)
 
+	c.initMetrics()
+
 	return c
 }
 
+// initMetrics initializes all metric label combinations with zero values.
+// This ensures metrics appear in Prometheus scrapes before any deployments occur.
+func (c *Collector) initMetrics() {
+	actions := []messages.Action{
+		messages.ActionSwitch,
+		messages.ActionBoot,
+		messages.ActionTest,
+		messages.ActionDryActivate,
+	}
+
+	// Initialize deployment counter for common status/action combinations
+	for _, action := range actions {
+		// Successful completions (no error code)
+		c.deploymentsTotal.WithLabelValues("completed", string(action), "")
+		// Failed deployments (no error code - from RecordDeploymentEnd)
+		c.deploymentsTotal.WithLabelValues("failed", string(action), "")
+	}
+
+	// Initialize histogram for all action/success combinations
+	for _, action := range actions {
+		c.deploymentDuration.WithLabelValues(string(action), "true")
+		c.deploymentDuration.WithLabelValues(string(action), "false")
+	}
+}
+
 // SetInfo sets the static instance metadata.
 func (c *Collector) SetInfo(hostname, tier, role, version string) {
 	c.info.WithLabelValues(hostname, tier, role, version).Set(1)