UDP drops should be a counter, not a gauge.

I think it makes sense for "drops" to sit alongside the queue length
gauges in prom stats because they are all neighbors in procfs (source
of these stats).

Moreover, in reading the commit log message for the original creating
work for udp_queues, I think there may have been some misreading or
confusion between the word "state" and the common short-form of
"stats" to mean "statistics".  The original author "chose the name
'udp_queue' instead of 'udpstat' as UDP has no state"; I believe that
'udpstat' might actually be the more appropriate name.

Signed-off-by: Chris Cleeland <[email protected]>

Author: cleeland

collector/udp_queues_linux.go

@@ -62,7 +62,7 @@ func (c *udpQueuesCollector) Update(ch chan<- prometheus.Metric) error {
 		ch <- prometheus.MustNewConstMetric(c.desc, prometheus.GaugeValue, float64(s4.TxQueueLength), "tx", "v4")
 		ch <- prometheus.MustNewConstMetric(c.desc, prometheus.GaugeValue, float64(s4.RxQueueLength), "rx", "v4")
 		if s4.Drops != nil {
-			ch <- prometheus.MustNewConstMetric(c.desc, prometheus.GaugeValue, float64(*s4.Drops), "drops", "v4")
+			ch <- prometheus.MustNewConstMetric(c.desc, prometheus.CounterValue, float64(*s4.Drops), "drops", "v4")
 		}
 	} else {
 		if errors.Is(errIPv4, os.ErrNotExist) {
@@ -77,7 +77,7 @@ func (c *udpQueuesCollector) Update(ch chan<- prometheus.Metric) error {
 		ch <- prometheus.MustNewConstMetric(c.desc, prometheus.GaugeValue, float64(s6.TxQueueLength), "tx", "v6")
 		ch <- prometheus.MustNewConstMetric(c.desc, prometheus.GaugeValue, float64(s6.RxQueueLength), "rx", "v6")
 		if s6.Drops != nil {
-			ch <- prometheus.MustNewConstMetric(c.desc, prometheus.GaugeValue, float64(*s6.Drops), "drops", "v6")
+			ch <- prometheus.MustNewConstMetric(c.desc, prometheus.CounterValue, float64(*s6.Drops), "drops", "v6")
 		}
 	} else {
 		if errors.Is(errIPv6, os.ErrNotExist) {