Fix GetAppClusteringInfo counter accounting

GetAppClusteringInfo() called IncrementRequestCounter() on both the
error and success paths. IncrementRequestCounter only bumps
totalRequests, not errors or successfulRequests, so every clustering
probe silently widened the gap between ovn_requests_total and
ovn_successful_requests_total + ovn_failed_requests_total.

Any dashboard computing a success ratio as
ovn_successful_requests_total / ovn_requests_total therefore saw an
artificial ceiling below 100% even when nothing was actually failing.

Use IncrementErrorCounter on the failure path and
IncrementSuccessCounter on the success path so the three request
counters stay consistent (total = successful + failed).

Signed-off-by: Michele Palazzi <sysdadmin@m1k.cloud>

Author: ironashram

pkg/ovn_exporter/ovn_exporter.go

@@ -1097,7 +1097,7 @@ func (e *Exporter) GatherMetrics() {
 						"system_id", e.Client.System.ID,
 						"error", err.Error(),
 					)
-					e.IncrementRequestCounter()
+					e.IncrementErrorCounter()
 					e.metrics = append(e.metrics, prometheus.MustNewConstMetric(
 						clusterEnabled,
 						prometheus.GaugeValue,
@@ -1106,7 +1106,7 @@ func (e *Exporter) GatherMetrics() {
 						component,
 					))
 				} else {
-					e.IncrementRequestCounter()
+					e.IncrementSuccessCounter()
 					isClusterEnabled = true
 					switch component {
 					case "ovsdb-server-southbound":