Add ingress gateway metrics

Signed-off-by: Raul Sevilla <[email protected]>
cloud-bulldozer · Jan 15, 2024 · 7caef38 · 7caef38
1 parent 292cf22
commit 7caef38
Show file tree

Hide file tree

Showing 4 changed files with 9 additions and 8 deletions.
diff --git a/pkg/config/config.go b/pkg/config/config.go
@@ -25,10 +25,9 @@ import (
 func (c *Config) UnmarshalYAML(unmarshal func(interface{}) error) error {
 	type ConfigDefaulted Config
 	defaultCfg := ConfigDefaulted{
-		Warmup:            false, // Disable warmup by default
-		RequestTimeout:    time.Second,
-		Procs:             1,
-		PrometheusMetrics: prometheusQueries,
+		Warmup:         false, // Disable warmup by default
+		RequestTimeout: time.Second,
+		Procs:          1,
 	}
 	if err := unmarshal(&defaultCfg); err != nil {
 		return err

diff --git a/pkg/config/types.go b/pkg/config/types.go
@@ -46,11 +46,9 @@ type Config struct {
 	Warmup bool `yaml:"warmup" json:"-"`
 	// RequestTimeout defines the tool request timeout
 	RequestTimeout time.Duration `yaml:"requestTimeout" json:"requestTimeout"`
-	// Prometheus metrics
-	PrometheusMetrics map[string]string `json:"-"`
 }
 
-var prometheusQueries = map[string]string{
+var PrometheusQueries = map[string]string{
 	"avg_cpu_usage_router_pods":           "avg(avg_over_time(sum(irate(container_cpu_usage_seconds_total{name!='', namespace='openshift-ingress', pod=~'router-default.+'}[2m])) by (pod)[ELAPSED:]))",
 	"avg_memory_usage_router_pods_bytes":  "avg(avg_over_time(sum(container_memory_working_set_bytes{name!='', namespace='openshift-ingress', pod=~'router-default.+'}) by (pod)[ELAPSED:]))",
 	"avg_cpu_usage_router_nodes":          "avg(avg_over_time(sum(irate(node_cpu_seconds_total{mode!~'idle|steal'}[2m]) and on (instance) label_replace(kube_pod_info{namespace='openshift-ingress'},'instance', '$1', 'node', '(.+)')) by (instance)[ELAPSED:]))",

diff --git a/pkg/runner/exec.go b/pkg/runner/exec.go
@@ -110,7 +110,7 @@ func runBenchmark(cfg config.Config, clusterMetadata tools.ClusterMetadata, p *p
 		timeouts += result.Timeouts
 		httpErrors += result.HTTPErrors
 		elapsed := fmt.Sprintf("%ds", int(time.Since(sampleTs).Seconds()))
-		for field, query := range cfg.PrometheusMetrics {
+		for field, query := range config.PrometheusQueries {
 			promQuery := strings.ReplaceAll(query, "ELAPSED", elapsed)
 			log.Debugf("Running query: %s", promQuery)
 			value, err := p.Query(promQuery, time.Time{}.UTC())

diff --git a/pkg/runner/runner.go b/pkg/runner/runner.go
@@ -94,6 +94,10 @@ func WithServiceMesh(enable bool, igNamespace string) OptsFunctions {
 	return func(r *Runner) {
 		r.serviceMesh = enable
 		r.igNamespace = igNamespace
+		config.PrometheusQueries["avg_cpu_usage_ingress_gateway_pods"] =
+			fmt.Sprintf("avg(avg_over_time(sum(irate(container_cpu_usage_seconds_total{name!='', namespace='%s', pod=~'istio-ingressgateway.+'}[2m])) by (pod)[ELAPSED:]))", igNamespace)
+		config.PrometheusQueries["avg_memory_usage_ingress_gateway_pods_bytes"] =
+			fmt.Sprintf("avg(avg_over_time(sum(container_memory_working_set_bytes{name!='', namespace='%s', pod=~'istio-ingressgateway.+'}) by (pod)[10m:]))", igNamespace)
 	}
 }