cgroups: Adjust metric names

Change the metric names to be more accurate. The memory.failcnt tracks
how often the kernel has tried to reclaim memory. While this maps
closely to the number of killed processes, they are not exactly the same
thing.

The cgroups cpu usage metrics are a running total. Add "total" at the
end of the metric to indicate this.

Changelog: changed

2 files changed, 15 insertions, 15 deletions

diff --git a/internal/cgroups/v1_linux.go b/internal/cgroups/v1_linux.go
index e4b72f28b..201d098fe 100644
--- a/internal/cgroups/v1_linux.go
+++ b/internal/cgroups/v1_linux.go
@@ -17,10 +17,10 @@ import (
 
 // CGroupV1Manager is the manager for cgroups v1
 type CGroupV1Manager struct {
-	cfg                         cgroupscfg.Config
-	hierarchy                   func() ([]cgroups.Subsystem, error)
-	memoryFailedTotal, cpuUsage *prometheus.GaugeVec
-	procs                       *prometheus.GaugeVec
+	cfg                                  cgroupscfg.Config
+	hierarchy                            func() ([]cgroups.Subsystem, error)
+	memoryReclaimAttemptsTotal, cpuUsage *prometheus.GaugeVec
+	procs                                *prometheus.GaugeVec
 }
 
 func newV1Manager(cfg cgroupscfg.Config) *CGroupV1Manager {
@@ -29,16 +29,16 @@ func newV1Manager(cfg cgroupscfg.Config) *CGroupV1Manager {
 		hierarchy: func() ([]cgroups.Subsystem, error) {
 			return defaultSubsystems(cfg.Mountpoint)
 		},
-		memoryFailedTotal: prometheus.NewGaugeVec(
+		memoryReclaimAttemptsTotal: prometheus.NewGaugeVec(
 			prometheus.GaugeOpts{
-				Name: "gitaly_cgroup_memory_failed_total",
+				Name: "gitaly_cgroup_memory_reclaim_attempts_total",
 				Help: "Number of memory usage hits limits",
 			},
 			[]string{"path"},
 		),
 		cpuUsage: prometheus.NewGaugeVec(
 			prometheus.GaugeOpts{
-				Name: "gitaly_cgroup_cpu_usage",
+				Name: "gitaly_cgroup_cpu_usage_total",
 				Help: "CPU Usage of Cgroup",
 			},
 			[]string{"path", "type"},
@@ -157,7 +157,7 @@ func (cg *CGroupV1Manager) Collect(ch chan<- prometheus.Metric) {
 		if metrics, err := control.Stat(); err != nil {
 			logger.WithError(err).Warn("unable to get cgroup stats")
 		} else {
-			memoryMetric := cg.memoryFailedTotal.WithLabelValues(repoPath)
+			memoryMetric := cg.memoryReclaimAttemptsTotal.WithLabelValues(repoPath)
 			memoryMetric.Set(float64(metrics.Memory.Usage.Failcnt))
 			ch <- memoryMetric
 
diff --git a/internal/cgroups/v1_linux_test.go b/internal/cgroups/v1_linux_test.go
index 844b06225..ede9155ba 100644
--- a/internal/cgroups/v1_linux_test.go
+++ b/internal/cgroups/v1_linux_test.go
@@ -188,13 +188,13 @@ func TestMetrics(t *testing.T) {
 
 	repoCgroupPath := filepath.Join(v1Manager1.currentProcessCgroup(), "repos-0")
 
-	expected := bytes.NewBufferString(fmt.Sprintf(`# HELP gitaly_cgroup_cpu_usage CPU Usage of Cgroup
-# TYPE gitaly_cgroup_cpu_usage gauge
-gitaly_cgroup_cpu_usage{path="%s",type="kernel"} 0
-gitaly_cgroup_cpu_usage{path="%s",type="user"} 0
-# HELP gitaly_cgroup_memory_failed_total Number of memory usage hits limits
-# TYPE gitaly_cgroup_memory_failed_total gauge
-gitaly_cgroup_memory_failed_total{path="%s"} 2
+	expected := bytes.NewBufferString(fmt.Sprintf(`# HELP gitaly_cgroup_cpu_usage_total CPU Usage of Cgroup
+# TYPE gitaly_cgroup_cpu_usage_total gauge
+gitaly_cgroup_cpu_usage_total{path="%s",type="kernel"} 0
+gitaly_cgroup_cpu_usage_total{path="%s",type="user"} 0
+# HELP gitaly_cgroup_memory_reclaim_attempts_total Number of memory usage hits limits
+# TYPE gitaly_cgroup_memory_reclaim_attempts_total gauge
+gitaly_cgroup_memory_reclaim_attempts_total{path="%s"} 2
 # HELP gitaly_cgroup_procs_total Total number of procs
 # TYPE gitaly_cgroup_procs_total gauge
 gitaly_cgroup_procs_total{path="%s",subsystem="cpu"} 1