MEDIUM: add prometheus metric haproxy_unable_to_sync_configuration

hdurand0710 · oktalz · commit b390cc7c01d6 · 2025-02-19T12:14:05.000Z
diff --git a/.aspell.yml b/.aspell.yml
@@ -36,3 +36,4 @@ allowed:
   - kubebuilder
   - cfg
   - optim
+  - prometheus
diff --git a/documentation/prometheus.md b/documentation/prometheus.md
@@ -26,6 +26,7 @@ On top of the prometheus provided metrics, we added these three ones:
 haproxy_reloads_total: The number of haproxy reloads partitioned by result (success/failure)
 haproxy_restarts_total: The number of haproxy restarts partitioned by result (success/failure)
 haproxy_runtime_socket_connections_total: The number of haproxy runtime socket connections partitioned by object (server/map) and result (success/failure)
+haproxy_unable_to_sync_configuration 1 = there's a pending haproxy configuration that is not valid so not applicable, 0 = haproxy configuration applied
 ```
 
 
@@ -46,4 +47,7 @@ haproxy_restarts_total{result="success"} 1
 # TYPE haproxy_runtime_socket_connections_total counter
 haproxy_runtime_socket_connections_total{object="map",result="success"} 4
 haproxy_runtime_socket_connections_total{object="server",result="success"} 50
+# HELP haproxy_unable_to_sync_configuration 1 = there's a pending haproxy configuration that is not valid so not applicable, 0 = haproxy configuration applied
+# TYPE haproxy_unable_to_sync_configuration gauge
+haproxy_unable_to_sync_configuration 1
 ```
diff --git a/pkg/controller/controller.go b/pkg/controller/controller.go
@@ -111,6 +111,7 @@ func (c *HAProxyController) Stop() {
 func (c *HAProxyController) updateHAProxy() {
 	var err error
 	logger.Trace("HAProxy config sync started")
+	c.prometheusMetricsManager.UnsetUnableSyncGauge()
 
 	err = c.haproxy.APIStartTransaction()
 	if err != nil {
@@ -183,6 +184,7 @@ func (c *HAProxyController) updateHAProxy() {
 
 	err = c.haproxy.APIFinalCommitTransaction()
 	if err != nil {
+		c.prometheusMetricsManager.SetUnableSyncGauge()
 		logger.Error("unable to Sync HAProxy configuration !!")
 		logger.Error(err)
 		rerun, errCfgSnippet := annotations.CheckBackendConfigSnippetError(err, c.haproxy.Env.CfgDir)
diff --git a/pkg/metrics/prometheus.go b/pkg/metrics/prometheus.go
@@ -15,6 +15,8 @@ const (
 )
 
 type PrometheusMetricsManager struct {
+	unableToSyncGauge prometheus.Gauge
+
 	// reload
 	reloadsCounterVec *prometheus.CounterVec
 
@@ -47,9 +49,15 @@ func New() PrometheusMetricsManager {
 			[]string{"object", "result"},
 		)
 
+		unableToSyncGauge := promauto.NewGauge(prometheus.GaugeOpts{
+			Name: "haproxy_unable_to_sync_configuration",
+			Help: "1 = there's a pending haproxy configuration that is not valid so not applicable, 0 = haproxy configuration applied",
+		})
+
 		pmm = PrometheusMetricsManager{
 			reloadsCounterVec:       reloadCounter,
 			runtimeSocketCounterVec: runtimeSocketCounter,
+			unableToSyncGauge:       unableToSyncGauge,
 		}
 	})
 	return pmm
@@ -70,3 +78,11 @@ func (pmm PrometheusMetricsManager) UpdateRuntimeMetrics(object string, err erro
 		pmm.runtimeSocketCounterVec.WithLabelValues(object, ResultSuccess).Inc()
 	}
 }
+
+func (pmm PrometheusMetricsManager) SetUnableSyncGauge() {
+	pmm.unableToSyncGauge.Set(float64(1))
+}
+
+func (pmm PrometheusMetricsManager) UnsetUnableSyncGauge() {
+	pmm.unableToSyncGauge.Set(float64(0))
+}
