fix(promql-to-scrape): fixed response type definitions

cloud/observability/promql-to-scrape/cmd/genconfig/main.go

@@ -45,9 +45,9 @@ func main() {
 	if err != nil {
 		log.Fatalf("Failed to pull metric names: %s", err)
 	}
-	fmt.Println(counters)
-	fmt.Println(gauges)
-	fmt.Println(histograms)
+	fmt.Println("counters: ", counters, "\n")
+	fmt.Println("gauges: ", gauges, "\n")
+	fmt.Println("histograms: ", histograms, "\n")
 
 	conf := internal.Config{}
 

cloud/observability/promql-to-scrape/examples/config.yaml

@@ -1,43 +1,43 @@
 metrics:
-    - metric_name: temporal_cloud_v0_frontend_service_error_count:rate1m
+    - metric_name: temporal_cloud_v0_frontend_service_error_count
       query: rate(temporal_cloud_v0_frontend_service_error_count[1m])
     - metric_name: temporal_cloud_v0_frontend_service_pending_requests
       query: temporal_cloud_v0_frontend_service_pending_requests
-    - metric_name: temporal_cloud_v0_frontend_service_request_count:rate1m
+    - metric_name: temporal_cloud_v0_frontend_service_request_count
       query: rate(temporal_cloud_v0_frontend_service_request_count[1m])
-    - metric_name: temporal_cloud_v0_poll_success_count:rate1m
+    - metric_name: temporal_cloud_v0_poll_success_count
       query: rate(temporal_cloud_v0_poll_success_count[1m])
-    - metric_name: temporal_cloud_v0_poll_success_sync_count:rate1m
+    - metric_name: temporal_cloud_v0_poll_success_sync_count
       query: rate(temporal_cloud_v0_poll_success_sync_count[1m])
-    - metric_name: temporal_cloud_v0_poll_timeout_count:rate1m
+    - metric_name: temporal_cloud_v0_poll_timeout_count
       query: rate(temporal_cloud_v0_poll_timeout_count[1m])
-    - metric_name: temporal_cloud_v0_resource_exhausted_error_count:rate1m
+    - metric_name: temporal_cloud_v0_resource_exhausted_error_count
       query: rate(temporal_cloud_v0_resource_exhausted_error_count[1m])
-    - metric_name: temporal_cloud_v0_schedule_action_success_count:rate1m
+    - metric_name: temporal_cloud_v0_schedule_action_success_count
       query: rate(temporal_cloud_v0_schedule_action_success_count[1m])
-    - metric_name: temporal_cloud_v0_schedule_buffer_overruns_count:rate1m
+    - metric_name: temporal_cloud_v0_schedule_buffer_overruns_count
       query: rate(temporal_cloud_v0_schedule_buffer_overruns_count[1m])
-    - metric_name: temporal_cloud_v0_schedule_missed_catchup_window_count:rate1m
+    - metric_name: temporal_cloud_v0_schedule_missed_catchup_window_count
       query: rate(temporal_cloud_v0_schedule_missed_catchup_window_count[1m])
-    - metric_name: temporal_cloud_v0_service_latency_bucket:histogram_quantile_p99_1m
+    - metric_name: temporal_cloud_v0_service_latency_bucket
       query: histogram_quantile(0.99, sum(rate(temporal_cloud_v0_service_latency_bucket[1m])) by (le, operation, temporal_namespace))
-    - metric_name: temporal_cloud_v0_service_latency_count:rate1m
+    - metric_name: temporal_cloud_v0_service_latency_count
       query: rate(temporal_cloud_v0_service_latency_count[1m])
-    - metric_name: temporal_cloud_v0_service_latency_sum:rate1m
+    - metric_name: temporal_cloud_v0_service_latency_sum
       query: rate(temporal_cloud_v0_service_latency_sum[1m])
-    - metric_name: temporal_cloud_v0_state_transition_count:rate1m
+    - metric_name: temporal_cloud_v0_state_transition_count
       query: rate(temporal_cloud_v0_state_transition_count[1m])
-    - metric_name: temporal_cloud_v0_total_action_count:rate1m
+    - metric_name: temporal_cloud_v0_total_action_count
       query: rate(temporal_cloud_v0_total_action_count[1m])
-    - metric_name: temporal_cloud_v0_workflow_cancel_count:rate1m
+    - metric_name: temporal_cloud_v0_workflow_cancel_count
       query: rate(temporal_cloud_v0_workflow_cancel_count[1m])
-    - metric_name: temporal_cloud_v0_workflow_continued_as_new_count:rate1m
+    - metric_name: temporal_cloud_v0_workflow_continued_as_new_count
       query: rate(temporal_cloud_v0_workflow_continued_as_new_count[1m])
-    - metric_name: temporal_cloud_v0_workflow_failed_count:rate1m
+    - metric_name: temporal_cloud_v0_workflow_failed_count
       query: rate(temporal_cloud_v0_workflow_failed_count[1m])
-    - metric_name: temporal_cloud_v0_workflow_success_count:rate1m
+    - metric_name: temporal_cloud_v0_workflow_success_count
       query: rate(temporal_cloud_v0_workflow_success_count[1m])
-    - metric_name: temporal_cloud_v0_workflow_terminate_count:rate1m
+    - metric_name: temporal_cloud_v0_workflow_terminate_count
       query: rate(temporal_cloud_v0_workflow_terminate_count[1m])
-    - metric_name: temporal_cloud_v0_workflow_timeout_count:rate1m
+    - metric_name: temporal_cloud_v0_workflow_timeout_count
       query: rate(temporal_cloud_v0_workflow_timeout_count[1m])

cloud/observability/promql-to-scrape/internal/client.go

@@ -71,9 +71,10 @@ func (c *APIClient) ListMetrics(metricPrefix string) ([]string, []string, []stri
 		if !strings.HasPrefix(string(v), metricPrefix) {
 			continue
 		}
-		if strings.HasSuffix(string(v), "_bucket") {
+		t := getMetricType(string(v))
+		if t == metricTypeHistogram {
 			histograms = append(histograms, string(v))
-		} else if strings.HasSuffix(string(v), "_count") || strings.HasSuffix(string(v), "_sum") {
+		} else if t == metricTypeCounter {
 			counts = append(counts, string(v))
 		} else {
 			gauges = append(gauges, string(v))

cloud/observability/promql-to-scrape/internal/metric.go

@@ -0,0 +1,18 @@
+package internal
+
+import "strings"
+
+const (
+	metricTypeHistogram = "histogram"
+	metricTypeCounter   = "count"
+	metricTypeGauge     = "gauge"
+)
+
+func getMetricType(v string) string {
+	if strings.HasSuffix(v, "_bucket") {
+		return metricTypeHistogram
+	} else if strings.HasSuffix(v, "_count") || strings.HasSuffix(v, "_sum") {
+		return metricTypeCounter
+	}
+	return metricTypeGauge
+}

cloud/observability/promql-to-scrape/internal/serialize.go

@@ -38,7 +38,7 @@ func SamplesToString(queriedMetrics map[string][]*model.Sample) string {
 		sb.WriteString("# TYPE ")
 		sb.WriteString(nameWithoutSuffix)
 		sb.WriteByte(' ')
-		sb.WriteString("gauge")
+		sb.WriteString(getMetricType(metricName))
 		sb.WriteByte('\n')
 
 		for _, s := range samples {

cloud/observability/promql-to-scrape/internal/server.go

@@ -46,7 +46,11 @@ func (s *PromToScrapeServer) metricsHandler(w http.ResponseWriter, r *http.Reque
 	s.RLock()
 	defer s.RUnlock()
 	if time.Since(s.lastSuccessfulTime) < 5*time.Minute {
-		fmt.Fprint(w, s.data)
+		_, err := fmt.Fprint(w, s.data)
+		if err != nil {
+			w.WriteHeader(http.StatusInternalServerError)
+			slog.Error("can't serve metrics", "error", err)
+		}
 	} else {
 		w.WriteHeader(http.StatusInternalServerError)
 		slog.Error("can't serve metrics", "error", "metrics queried are stale (more than 5 minutes old)")
@@ -71,7 +75,7 @@ func (s *PromToScrapeServer) run() string {
 //
 //	keep the objects returned from the query, or convert them into something a bit more ergonomic
 //	and create ConstMetrics with the prometheus client. I happened to have the code lying around for working
-//	with model.Sample, but the CosntMetrics route is probably more idiomatic and safe.
+//	with model.Sample, but the ConstMetrics route is probably more idiomatic and safe.
 func (s *PromToScrapeServer) queryMetrics() {
 	start := time.Now()
 	queriedMetrics, err := QueryMetrics(s.conf, s.client)
@@ -88,5 +92,6 @@ func (s *PromToScrapeServer) queryMetrics() {
 
 // Start runs the embedded http.Server.
 func (s *PromToScrapeServer) Start() error {
+	slog.Info("listening on", "addr", s.server.Addr)
 	return s.server.ListenAndServe()
 }