Initialise reason="connect-timeout" series for cache `operation_fai…

…lures_total` metric (#545) * Initialise `reason="connect-timeout"` series for memcached `operation_failures_total` metric * Add changelog entry
grafana · Jul 12, 2024 · b834d6b · b834d6b
1 parent 6c2188e
commit b834d6b
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 0 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -247,3 +247,4 @@
 * [BUGFIX] middleware: fix issue where applications that used the httpgrpc tracing middleware would generate duplicate spans for incoming HTTP requests. #451
 * [BUGFIX] httpgrpc: store headers in canonical form when converting from gRPC to HTTP. #518
 * [BUGFIX] Memcached: Don't truncate sub-second TTLs to 0 which results in them being cached forever. #530 
+* [BUGFIX] Cache: initialise the `operation_failures_total{reason="connect-timeout"}` metric to 0 for each cache operation type on startup. #545
diff --git a/cache/client.go b/cache/client.go
@@ -86,6 +86,7 @@ func newClientMetrics(reg prometheus.Registerer) *clientMetrics {
 		Help: "Total number of operations against cache that failed.",
 	}, []string{"operation", "reason"})
 	for _, op := range []string{opGetMulti, opSet, opDelete, opIncrement, opFlush, opTouch, opCompareAndSwap} {
+		cm.failures.WithLabelValues(op, reasonConnectTimeout)
 		cm.failures.WithLabelValues(op, reasonTimeout)
 		cm.failures.WithLabelValues(op, reasonMalformedKey)
 		cm.failures.WithLabelValues(op, reasonServerError)