feat!: rename metrics and service (#730)

## This PR

Fixes #712 

Refer the following table for metrics renames and their units,


 |  Old | New  | Unit |
|---|---|---|
| http_request_duration_seconds  |  http.server.duration | s |
| http_response_size_bytes | http.server.response.size | By |
| http_requests_inflight |http.server.active_requests | {request} |
| impressions | feature_flag.flagd.impression| {impression} | 
| reasons | feature_flag.flagd.evaluation.reason| {reason} |

### Validation

This change was validated against an OTel setup. Consider following
screen captures,

![Screenshot 2023-07-04 at 1 16 02
PM](https://github.com/open-feature/flagd/assets/8186721/5aba3103-e3fc-42a7-a396-c78999d22b05)

![Screenshot 2023-07-04 at 1 15 41
PM](https://github.com/open-feature/flagd/assets/8186721/0ea196f8-fcb4-4e8c-86eb-c6774de3e2f3)

![Screenshot 2023-07-04 at 1 14 53
PM](https://github.com/open-feature/flagd/assets/8186721/c285d78d-525a-4e40-bfe2-895a85d96cd6)

![Screenshot 2023-07-04 at 1 14 40
PM](https://github.com/open-feature/flagd/assets/8186721/e826ed8c-6056-4fd4-b206-d76592d8d484)

![Screenshot 2023-07-04 at 1 14 18
PM](https://github.com/open-feature/flagd/assets/8186721/30c959af-197b-405c-b303-c4dd59807311)

Signed-off-by: Kavindu Dodanduwa <kavindudodanduwa@gmail.com>

core/pkg/runtime/from_config.go

@@ -33,7 +33,7 @@ const (
  syncProviderGrpc = "grpc"
  syncProviderKubernetes = "kubernetes"
  syncProviderHTTP = "http"
- svcName = "openfeature/flagd"
+ svcName = "flagd"
 )
 
 var (

core/pkg/service/sync/sync_metrics.go

@@ -10,7 +10,7 @@ import (
 )
 
 const (
- serviceName = "openfeature/flagd-proxy"
+ serviceName = "flagd-proxy"
 )
 
 func (s *Server) captureMetrics() error {

core/pkg/telemetry/metrics.go

@@ -15,13 +15,16 @@ import (
 )
 
 const (
- requestDurationName = "http_request_duration_seconds"
- responseSizeName = "http_response_size_bytes"
- FlagdProviderName = "flagd"
- FeatureFlagReasonKeyName = "feature_flag.reason"
- ExceptionTypeKeyName = "exception.type"
- FeatureFlagReasonKey = attribute.Key(FeatureFlagReasonKeyName)
- ExceptionTypeKey = attribute.Key(ExceptionTypeKeyName)
+ ProviderName = "flagd"
+
+ FeatureFlagReasonKey = attribute.Key("feature_flag.reason")
+ ExceptionTypeKey = attribute.Key("ExceptionTypeKeyName")
+
+ httpRequestDurationMetric = "http.server.duration"
+ httpResponseSizeMetric = "http.server.response.size"
+ httpActiveRequestsMetric = "http.server.active_requests"
+ impressionMetric = "feature_flag." + ProviderName + ".impression"
+ reasonMetric = "feature_flag." + ProviderName + ".evaluation.reason"
 )
 
 type MetricsRecorder struct {
@@ -72,7 +75,7 @@ func (r MetricsRecorder) Impressions(ctx context.Context, reason, variant, key s
 
 func (r MetricsRecorder) Reasons(ctx context.Context, key string, reason string, err error) {
  attrs := []attribute.KeyValue{
- semconv.FeatureFlagProviderName(FlagdProviderName),
+ semconv.FeatureFlagProviderName(ProviderName),
  FeatureFlagReason(reason),
  }
  if err == nil {
@@ -116,9 +119,9 @@ func NewOTelRecorder(exporter msdk.Reader, resource *resource.Resource, serviceN
  provider := msdk.NewMeterProvider(
  msdk.WithReader(exporter),
  // for the request duration metric we use the default bucket size which are tailored for response time in seconds
- msdk.WithView(getDurationView(requestDurationName, serviceName, prometheus.DefBuckets)),
+ msdk.WithView(getDurationView(httpRequestDurationMetric, serviceName, prometheus.DefBuckets)),
  // for response size we want 8 exponential bucket starting from 100 Bytes
- msdk.WithView(getDurationView(responseSizeName, serviceName, prometheus.ExponentialBuckets(100, 10, 8))),
+ msdk.WithView(getDurationView(httpResponseSizeMetric, serviceName, prometheus.ExponentialBuckets(100, 10, 8))),
  // set entity producing telemetry
  msdk.WithResource(resource),
  )
@@ -127,25 +130,29 @@ func NewOTelRecorder(exporter msdk.Reader, resource *resource.Resource, serviceN
 
  // we can ignore errors from OpenTelemetry since they could occur if we select the wrong aggregator
  hduration, _ := meter.Float64Histogram(
- requestDurationName,
+ httpRequestDurationMetric,
  metric.WithDescription("The latency of the HTTP requests"),
+ metric.WithUnit("s"),
  )
  hsize, _ := meter.Float64Histogram(
- responseSizeName,
+ httpResponseSizeMetric,
  metric.WithDescription("The size of the HTTP responses"),
  metric.WithUnit("By"),
  )
  reqCounter, _ := meter.Int64UpDownCounter(
- "http_requests_inflight",
+ httpActiveRequestsMetric,
  metric.WithDescription("The number of inflight requests being handled at the same time"),
+ metric.WithUnit("{request}"),
  )
  impressions, _ := meter.Int64Counter(
- "impressions",
+ impressionMetric,
  metric.WithDescription("The number of evaluations for a given flag"),
+ metric.WithUnit("{impression}"),
  )
  reasons, _ := meter.Int64Counter(
- "reasons",
+ reasonMetric,
  metric.WithDescription("The number of evaluations for a given reason"),
+ metric.WithUnit("{reason}"),
  )
  return &MetricsRecorder{
  httpRequestDurHistogram: hduration,