Miscellanous compliance fixes

CHANGELOG.md

@@ -27,11 +27,19 @@ versioning](https://go.dev/doc/modules/version-numbers).
 - The Go generator now removes all the old defer statements in function bodies before re-adding
   only the necessary ones. This means calling `go generate` on a file that has no annotation
   at all effectively cleans up the whole file from autometrics.
-
+- All metrics now have a `service_name` label, which can either be compiled in `Init` call, or
+  filled at runtime from environment variables (in order of precedence):
+  + `AUTOMETRICS_SERVICE_NAME`
+  + `OTEL_SERVICE_NAME`
+
 ### Changed
 
 - Instead of returning an error when the go generator does not find the autometrics import
   in a file, it will add the needed import itself in the file.
+- Function calls metric has been renamed from `function_calls_count_total` to `function_calls_total`
+- Function calls duration histogram has been renamed from `function_calls_duration` 
+  to `function_calls_duration_seconds`
+- Function caller label has been split from `caller` to `caller_function` and `caller_label`
 
 ### Fixed
 

README.md

@@ -96,14 +96,19 @@ And then in your main function initialize the metrics
 	autometrics.Init(
 		nil,
 		autometrics.DefBuckets,
-		autometrics.BuildInfo{Version: "0.4.0", Commit: "anySHA", Branch: ""},
+		autometrics.BuildInfo{Version: "0.4.0", Commit: "anySHA", Branch: "", Service: "myApp"},
 	)
 ```
 
 Everything in `BuildInfo` is optional. It will add relevant information on the
 metrics for better intelligence. You can use any string variable whose value is
 injected at build time by `ldflags` for example, or use environment variables.
 
+> **Note**
+> Instead of hardcoding the service in the code, you can simply have environment variables set to fill the "Service" name.
+`AUTOMETRICS_SERVICE_NAME` will be used if set, otherwise `OTEL_SERVICE_NAME` will be attempted (so OpenTelemetry
+compatibility comes out of the box).
+
 ### 3. Add directives for each function you want to instrument
 
 > **Warning**
@@ -235,7 +240,7 @@ func main() {
 	autometrics.Init(
 		nil,
 		autometrics.DefBuckets,
-		autometrics.BuildInfo{Version: "0.4.0", Commit: "anySHA", Branch: ""},
+		autometrics.BuildInfo{Version: "0.4.0", Commit: "anySHA", Branch: "", Service: "myApp"},
 	)
 	http.Handle("/metrics", promhttp.Handler())
 }
@@ -345,7 +350,7 @@ metric. You can use the name of the application or its version for example
 -		nil,
 +		"myApp/v2/prod",
 		autometrics.DefBuckets,
-		autometrics.BuildInfo{ Version: "2.1.37", Commit: "anySHA", Branch: "" },
+		autometrics.BuildInfo{ Version: "2.1.37", Commit: "anySHA", Branch: "", Service: "myApp" },
 	)
 ```
 

docker-compose.open-telemetry-example.yaml

@@ -51,6 +51,8 @@ services:
     build:
       context: .
       dockerfile: examples/otel/Dockerfile
+    environment:
+      AUTOMETRICS_SERVICE_NAME: open-telemetry-autometrics-demo
     container_name: web-server-otel
     restart: unless-stopped
     expose:

docker-compose.prometheus-example.yaml

@@ -52,6 +52,8 @@ services:
     build:
       context: .
       dockerfile: examples/web/Dockerfile
+    environment:
+      AUTOMETRICS_SERVICE_NAME: prometheus-autometrics-demo
     container_name: web-server-prom
     restart: unless-stopped
     expose:

examples/otel/scripts/poll_server

@@ -6,11 +6,11 @@ TARGET_HOST="${TARGET_HOST:-localhost}"
 
 while true
 do
-    if [ "$(($RANDOM % 2))" == "0" ]; then
-        curl "http://${TARGET_HOST}:62086/random-error"
+    if [ "$(($RANDOM % 3))" == "0" ]; then
+        curl -s "http://${TARGET_HOST}:62086/random-error"
     fi
     if [ "$(($RANDOM % 4))" == "0" ]; then
-        curl "http://${TARGET_HOST}:62086/"
+        curl -s "http://${TARGET_HOST}:62086/"
     fi
-    sleep 0.2
+    sleep 0.05
 done

examples/web/cmd/main.go

@@ -89,21 +89,22 @@ func main() {
 // [Request Rate Callee]: http://localhost:9090/graph?g0.expr=%23+Rate+of+function+calls+emanating+from+%60indexHandler%60+function+per+second%2C+averaged+over+5+minute+windows%0A%0Asum+by+%28function%2C+module%2C+version%2C+commit%29+%28rate%28function_calls_count%7Bcaller%3D%22main.indexHandler%22%7D%5B5m%5D%29+%2A+on+%28instance%2C+job%29+group_left%28version%2C+commit%29+last_over_time%28build_info%5B1s%5D%29%29&g0.tab=0
 // [Error Ratio Callee]: http://localhost:9090/graph?g0.expr=%23+Percentage+of+function+emanating+from+%60indexHandler%60+function+that+return+errors%2C+averaged+over+5+minute+windows%0A%0A%28sum+by+%28function%2C+module%2C+version%2C+commit%29+%28rate%28function_calls_count%7Bcaller%3D%22main.indexHandler%22%2Cresult%3D%22error%22%7D%5B5m%5D%29+%2A+on+%28instance%2C+job%29+group_left%28version%2C+commit%29+last_over_time%28build_info%5B1s%5D%29%29%29+%2F+%28sum+by+%28function%2C+module%2C+version%2C+commit%29+%28rate%28function_calls_count%7Bcaller%3D%22main.indexHandler%22%7D%5B5m%5D%29+%2A+on+%28instance%2C+job%29+group_left%28version%2C+commit%29+last_over_time%28build_info%5B1s%5D%29%29%29&g0.tab=0
 //
-//autometrics:inst --slo "API" --latency-target 99 --latency-ms 100
+//autometrics:inst --slo "API" --latency-target 99 --latency-ms 5
 func indexHandler(w http.ResponseWriter, r *http.Request) error {
 	defer autometrics.Instrument(autometrics.PreInstrument(autometrics.NewContext(
 		r.Context(),
 		autometrics.WithConcurrentCalls(true),
 		autometrics.WithCallerName(true),
 		autometrics.WithSloName("API"),
-		autometrics.WithAlertLatency(100000000*time.Nanosecond, 99),
+		autometrics.WithAlertLatency(5000000*time.Nanosecond, 99),
 	)), nil) //autometrics:defer
 
-	time.Sleep(time.Duration(rand.Intn(200)) * time.Millisecond)
+	msSleep := rand.Intn(200)
+	time.Sleep(time.Duration(msSleep) * time.Millisecond)
 
-	fmt.Fprintf(w, "Hello, World!\n")
+	_, err := fmt.Fprintf(w, "Slept %v ms\n", msSleep)
 
-	return nil
+	return err
 }
 
 var handlerError = errors.New("failed to handle request")

examples/web/cmd/main.go.orig

@@ -61,14 +61,15 @@ func main() {
 //
 // It always succeeds and says hello.
 //
-//autometrics:inst --slo "API" --latency-target 99 --latency-ms 100
+//autometrics:inst --slo "API" --latency-target 99 --latency-ms 5
 func indexHandler(w http.ResponseWriter, r *http.Request) error {
 
-	time.Sleep(time.Duration(rand.Intn(200)) * time.Millisecond)
+	msSleep :=  rand.Intn(200)
+	time.Sleep(time.Duration(msSleep) * time.Millisecond)
 
-	fmt.Fprintf(w, "Hello, World!\n")
+	_, err := fmt.Fprintf(w, "Slept %v ms\n", msSleep)
 
-	return nil
+	return err
 }
 
 var handlerError = errors.New("failed to handle request")

examples/web/scripts/poll_server

@@ -6,11 +6,11 @@ TARGET_HOST="${TARGET_HOST:-localhost}"
 
 while true
 do
-    if [ "$(($RANDOM % 2))" == "0" ]; then
-        curl "http://${TARGET_HOST}:62086/random-error"
+    if [ "$(($RANDOM % 3))" == "0" ]; then
+        curl -s "http://${TARGET_HOST}:62086/random-error"
     fi
     if [ "$(($RANDOM % 4))" == "0" ]; then
-        curl "http://${TARGET_HOST}:62086/"
+        curl -s "http://${TARGET_HOST}:62086/"
     fi
-    sleep 0.2
+    sleep 0.05
 done

internal/autometrics/prometheus_link_gen.go

@@ -55,9 +55,10 @@ func addBuildInfoLabels() string {
 }
 
 func requestRateQuery(counterName, labelKey, labelValue string) string {
-	return fmt.Sprintf("sum by (%s, %s, %s, %s) (rate(%s{%s=\"%s\"}[5m]) %s)",
+	return fmt.Sprintf("sum by (%s, %s, %s, %s, %s) (rate(%s{%s=\"%s\"}[5m]) %s)",
 		prometheus.FunctionLabel,
 		prometheus.ModuleLabel,
+		prometheus.ServiceNameLabel,
 		prometheus.VersionLabel,
 		prometheus.CommitLabel,
 		counterName,
@@ -68,9 +69,10 @@ func requestRateQuery(counterName, labelKey, labelValue string) string {
 }
 
 func errorRatioQuery(counterName, labelKey, labelValue string) string {
-	return fmt.Sprintf("(sum by (%s, %s, %s, %s) (rate(%s{%s=\"%s\",%s=\"error\"}[5m]) %s)) / (%s)",
+	return fmt.Sprintf("(sum by (%s, %s, %s, %s, %s) (rate(%s{%s=\"%s\",%s=\"error\"}[5m]) %s)) / (%s)",
 		prometheus.FunctionLabel,
 		prometheus.ModuleLabel,
+		prometheus.ServiceNameLabel,
 		prometheus.VersionLabel,
 		prometheus.CommitLabel,
 		counterName,
@@ -83,9 +85,10 @@ func errorRatioQuery(counterName, labelKey, labelValue string) string {
 }
 
 func latencyQuery(bucketName, labelKey, labelValue string) string {
-	latency := fmt.Sprintf("sum by (le, %s, %s, %s, %s) (rate(%s_bucket{%s=\"%s\"}[5m]) %s)",
+	latency := fmt.Sprintf("sum by (le, %s, %s, %s, %s, %s) (rate(%s_bucket{%s=\"%s\"}[5m]) %s)",
 		prometheus.FunctionLabel,
 		prometheus.ModuleLabel,
+		prometheus.ServiceNameLabel,
 		prometheus.VersionLabel,
 		prometheus.CommitLabel,
 		bucketName,
@@ -103,9 +106,10 @@ func latencyQuery(bucketName, labelKey, labelValue string) string {
 }
 
 func concurrentCallsQuery(gaugeName, labelKey, labelValue string) string {
-	return fmt.Sprintf("sum by (%s, %s, %s, %s) (%s{%s=\"%s\"} %s)",
+	return fmt.Sprintf("sum by (%s, %s, %s, %s, %s) (%s{%s=\"%s\"} %s)",
 		prometheus.FunctionLabel,
 		prometheus.ModuleLabel,
+		prometheus.ServiceNameLabel,
 		prometheus.VersionLabel,
 		prometheus.CommitLabel,
 		gaugeName,
@@ -119,11 +123,11 @@ func (p Prometheus) GenerateAutometricsComment(ctx GeneratorContext, funcName, m
 	requestRateUrl := p.makePrometheusUrl(
 		requestRateQuery(prometheus.FunctionCallsCountName, prometheus.FunctionLabel, funcName), fmt.Sprintf("Rate of calls to the `%s` function per second, averaged over 5 minute windows", funcName))
 	calleeRequestRateUrl := p.makePrometheusUrl(
-		requestRateQuery(prometheus.FunctionCallsCountName, prometheus.CallerLabel, fmt.Sprintf("%s.%s", moduleName, funcName)), fmt.Sprintf("Rate of function calls emanating from `%s` function per second, averaged over 5 minute windows", funcName))
+		requestRateQuery(prometheus.FunctionCallsCountName, prometheus.CallerFunctionLabel, funcName), fmt.Sprintf("Rate of function calls emanating from `%s` function per second, averaged over 5 minute windows", funcName))
 	errorRatioUrl := p.makePrometheusUrl(
 		errorRatioQuery(prometheus.FunctionCallsCountName, prometheus.FunctionLabel, funcName), fmt.Sprintf("Percentage of calls to the `%s` function that return errors, averaged over 5 minute windows", funcName))
 	calleeErrorRatioUrl := p.makePrometheusUrl(
-		errorRatioQuery(prometheus.FunctionCallsCountName, prometheus.CallerLabel, fmt.Sprintf("%s.%s", moduleName, funcName)), fmt.Sprintf("Percentage of function emanating from `%s` function that return errors, averaged over 5 minute windows", funcName))
+		errorRatioQuery(prometheus.FunctionCallsCountName, prometheus.CallerFunctionLabel, funcName), fmt.Sprintf("Percentage of function emanating from `%s` function that return errors, averaged over 5 minute windows", funcName))
 	latencyUrl := p.makePrometheusUrl(
 		latencyQuery(prometheus.FunctionCallsDurationName, prometheus.FunctionLabel, funcName), fmt.Sprintf("95th and 99th percentile latencies (in seconds) for the `%s` function", funcName))
 	concurrentCallsUrl := p.makePrometheusUrl(