feat(go): expose runtime ingest metrics

This commit is contained in:
lingniu
2026-07-02 19:06:24 +08:00
parent 1dfd540700
commit 896e841902
18 changed files with 602 additions and 28 deletions

View File

@@ -17,6 +17,7 @@ import (
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/envelope"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/eventbus"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/identity"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/metrics"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/protocol/yutongmqtt"
)
@@ -38,6 +39,7 @@ type MQTTClientConfig struct {
Sink eventbus.Sink
Resolver identity.Resolver
Logger *slog.Logger
Metrics *metrics.Registry
}
type MQTTClient struct {
@@ -208,14 +210,41 @@ func (c *MQTTClient) handleMessage(ctx context.Context, topic string, payload []
env = resolved
}
}
c.recordFrameMetric(env.ParseStatus)
if err := c.cfg.Sink.PublishRaw(messageCtx, env); err != nil {
c.recordPublishMetric("raw", "error")
c.cfg.Logger.Error("publish mqtt raw failed", "topic", topic, "event_id", env.StableEventID(), "error", err)
return
}
c.recordPublishMetric("raw", "ok")
if env.ParseStatus == envelope.ParseBadFrame {
return
}
if err := c.cfg.Sink.PublishUnified(messageCtx, env); err != nil {
c.recordPublishMetric("unified", "error")
c.cfg.Logger.Error("publish mqtt unified failed", "topic", topic, "event_id", env.StableEventID(), "error", err)
return
}
c.recordPublishMetric("unified", "ok")
}
func (c *MQTTClient) recordFrameMetric(status envelope.ParseStatus) {
if c.cfg.Metrics == nil {
return
}
c.cfg.Metrics.IncCounter("vehicle_gateway_frames_total", metrics.Labels{
"protocol": string(envelope.ProtocolYutongMQTT),
"status": string(status),
})
}
func (c *MQTTClient) recordPublishMetric(kind string, status string) {
if c.cfg.Metrics == nil {
return
}
c.cfg.Metrics.IncCounter("vehicle_gateway_publish_total", metrics.Labels{
"protocol": string(envelope.ProtocolYutongMQTT),
"kind": kind,
"status": status,
})
}

View File

@@ -11,10 +11,12 @@ import (
"math/big"
"os"
"path/filepath"
"strings"
"testing"
"time"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/envelope"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/metrics"
)
func TestMQTTClientHandleMessagePublishesRawAndUnified(t *testing.T) {
@@ -45,6 +47,39 @@ func TestMQTTClientHandleMessagePublishesRawAndUnified(t *testing.T) {
}
}
func TestMQTTClientRecordsMessageMetrics(t *testing.T) {
registry := metrics.NewRegistry()
client, err := NewMQTTClient(MQTTClientConfig{
EndpointName: "endpoint-a",
Broker: "tcp://127.0.0.1:1883",
ClientID: "test-client",
Topics: []string{"/ytforward/shln/+"},
Sink: &recordingSink{},
Logger: slog.New(slog.NewTextHandler(testWriter{t: t}, nil)),
Metrics: registry,
})
if err != nil {
t.Fatalf("NewMQTTClient() error = %v", err)
}
client.handleMessage(context.Background(), "/ytforward/shln/dev1", []byte(`{
"device":"LTEST000000000001",
"time":"20260413100000",
"data":{"METER_SPEED":52.3,"TOTAL_MILEAGE":123456.7}
}`))
text := registry.Render()
for _, want := range []string{
`vehicle_gateway_frames_total{protocol="YUTONG_MQTT",status="OK"} 1`,
`vehicle_gateway_publish_total{kind="raw",protocol="YUTONG_MQTT",status="ok"} 1`,
`vehicle_gateway_publish_total{kind="unified",protocol="YUTONG_MQTT",status="ok"} 1`,
} {
if !strings.Contains(text, want) {
t.Fatalf("metrics missing %s:\n%s", want, text)
}
}
}
func TestMQTTClientHandleBadPayloadPublishesOnlyRaw(t *testing.T) {
sink := &recordingSink{}
client, err := NewMQTTClient(MQTTClientConfig{

View File

@@ -15,6 +15,7 @@ import (
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/envelope"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/eventbus"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/identity"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/metrics"
)
type FrameExtractor func([]byte) (frames [][]byte, remainder []byte, err error)
@@ -36,6 +37,7 @@ type TCPServer struct {
sink eventbus.Sink
resolver identity.Resolver
logger *slog.Logger
metrics *metrics.Registry
readBufferSize int
idleTimeout time.Duration
maxConnections int
@@ -46,6 +48,7 @@ type TCPServerConfig struct {
Sink eventbus.Sink
Resolver identity.Resolver
Logger *slog.Logger
Metrics *metrics.Registry
ReadBufferSize int
IdleTimeout time.Duration
MaxConnections int
@@ -89,6 +92,7 @@ func NewTCPServer(cfg TCPServerConfig) (*TCPServer, error) {
sink: cfg.Sink,
resolver: cfg.Resolver,
logger: cfg.Logger,
metrics: cfg.Metrics,
readBufferSize: cfg.ReadBufferSize,
idleTimeout: cfg.IdleTimeout,
maxConnections: cfg.MaxConnections,
@@ -209,18 +213,23 @@ func (s *TCPServer) handleFrame(ctx context.Context, conn net.Conn, raw []byte,
env = resolved
}
}
s.recordFrameMetric(env.ParseStatus)
if err := s.sink.PublishRaw(frameCtx, env); err != nil {
s.recordPublishMetric("raw", "error")
s.logger.Error("publish raw failed", "protocol", s.protocol.Protocol, "event_id", env.StableEventID(), "error", err)
return
}
s.recordPublishMetric("raw", "ok")
if env.ParseStatus == envelope.ParseBadFrame {
return
}
if err := s.sink.PublishUnified(frameCtx, env); err != nil {
s.recordPublishMetric("unified", "error")
s.logger.Error("publish unified failed", "protocol", s.protocol.Protocol, "event_id", env.StableEventID(), "error", err)
return
}
s.recordPublishMetric("unified", "ok")
if s.protocol.Respond == nil {
return
}
@@ -238,6 +247,27 @@ func (s *TCPServer) handleFrame(ctx context.Context, conn net.Conn, raw []byte,
}
}
func (s *TCPServer) recordFrameMetric(status envelope.ParseStatus) {
if s.metrics == nil {
return
}
s.metrics.IncCounter("vehicle_gateway_frames_total", metrics.Labels{
"protocol": string(s.protocol.Protocol),
"status": string(status),
})
}
func (s *TCPServer) recordPublishMetric(kind string, status string) {
if s.metrics == nil {
return
}
s.metrics.IncCounter("vehicle_gateway_publish_total", metrics.Labels{
"protocol": string(s.protocol.Protocol),
"kind": kind,
"status": status,
})
}
func (p TCPProtocol) String() string {
return fmt.Sprintf("%s@%s", p.Protocol, p.Addr)
}

View File

@@ -6,10 +6,12 @@ import (
"io"
"log/slog"
"net"
"strings"
"testing"
"time"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/envelope"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/metrics"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/protocol/gb32960"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/protocol/jt808"
)
@@ -45,6 +47,47 @@ func TestTCPServerPublishesGoodFrameToRawAndUnified(t *testing.T) {
}
}
func TestTCPServerRecordsFrameMetrics(t *testing.T) {
registry := metrics.NewRegistry()
server, err := NewTCPServer(TCPServerConfig{
Protocol: TCPProtocol{
Protocol: envelope.ProtocolJT808,
Addr: ":0",
Extract: jt808.ExtractFrames,
Parse: func(_ []byte, receivedAtMS int64, sourceEndpoint string) (envelope.FrameEnvelope, error) {
return envelope.FrameEnvelope{
Protocol: envelope.ProtocolJT808,
MessageID: "0x0200",
Phone: "13307795425",
SourceEndpoint: sourceEndpoint,
ReceivedAtMS: receivedAtMS,
EventTimeMS: receivedAtMS,
ParseStatus: envelope.ParseOK,
}, nil
},
},
Sink: &recordingSink{},
Logger: slog.New(slog.NewTextHandler(testWriter{t: t}, nil)),
Metrics: registry,
})
if err != nil {
t.Fatalf("NewTCPServer() error = %v", err)
}
server.handleFrame(context.Background(), nil, []byte{0x01}, "127.0.0.1:808")
text := registry.Render()
for _, want := range []string{
`vehicle_gateway_frames_total{protocol="JT808",status="OK"} 1`,
`vehicle_gateway_publish_total{kind="raw",protocol="JT808",status="ok"} 1`,
`vehicle_gateway_publish_total{kind="unified",protocol="JT808",status="ok"} 1`,
} {
if !strings.Contains(text, want) {
t.Fatalf("metrics missing %s:\n%s", want, text)
}
}
}
func TestTCPServerPublishesBadFrameOnlyToRaw(t *testing.T) {
good := buildGBFrame(0x02, 0xfe, "LNBSCB3D4R1234567", nil)
good[len(good)-1] ^= 0xff

View File

@@ -7,6 +7,8 @@ import (
"net/http"
"strings"
"time"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/metrics"
)
type Check struct {
@@ -32,22 +34,25 @@ func NewHandler(service string, checks []Check) *Handler {
return &Handler{service: service, checks: checks}
}
func NewMux(service string, checks []Check) *http.ServeMux {
func NewMux(service string, checks []Check, registry *metrics.Registry) *http.ServeMux {
mux := http.NewServeMux()
handler := NewHandler(service, checks)
mux.Handle("/healthz", handler)
mux.Handle("/readyz", handler)
if registry != nil {
mux.Handle("/metrics", metrics.NewHandler(registry))
}
return mux
}
func NewServer(addr string, service string, checks []Check) *http.Server {
func NewServer(addr string, service string, checks []Check, registry *metrics.Registry) *http.Server {
addr = strings.TrimSpace(addr)
if addr == "" {
return nil
}
return &http.Server{
Addr: addr,
Handler: NewMux(service, checks),
Handler: NewMux(service, checks, registry),
ReadHeaderTimeout: 5 * time.Second,
}
}

View File

@@ -7,6 +7,8 @@ import (
"net/http/httptest"
"strings"
"testing"
"lingniu-vehicle-ingest/go/vehicle-gateway/internal/metrics"
)
func TestHandlerReturnsHealthyWhenAllChecksPass(t *testing.T) {
@@ -63,10 +65,12 @@ func TestHandlerRejectsUnknownPath(t *testing.T) {
}
func TestNewMuxRegistersHealthAndReadinessRoutes(t *testing.T) {
registry := metrics.NewRegistry()
registry.IncCounter("vehicle_test_total", metrics.Labels{"service": "stat"})
mux := NewMux("vehicle-stat-writer", []Check{
{Name: "mysql", Check: func(context.Context) error { return nil }},
})
for _, path := range []string{"/healthz", "/readyz"} {
}, registry)
for _, path := range []string{"/healthz", "/readyz", "/metrics"} {
request := httptest.NewRequest(http.MethodGet, path, nil)
response := httptest.NewRecorder()
@@ -79,13 +83,13 @@ func TestNewMuxRegistersHealthAndReadinessRoutes(t *testing.T) {
}
func TestNewServerReturnsNilWhenAddressIsEmpty(t *testing.T) {
if server := NewServer("", "vehicle-gateway", nil); server != nil {
if server := NewServer("", "vehicle-gateway", nil, nil); server != nil {
t.Fatalf("server = %#v, want nil", server)
}
}
func TestNewServerBuildsConfiguredHTTPServer(t *testing.T) {
server := NewServer(":20290", "vehicle-gateway", nil)
server := NewServer(":20290", "vehicle-gateway", nil, nil)
if server == nil {
t.Fatal("server is nil")
}

View File

@@ -0,0 +1,141 @@
package metrics
import (
"fmt"
"net/http"
"sort"
"strings"
"sync"
)
type Labels map[string]string
type Registry struct {
mu sync.RWMutex
counters map[string]map[string]float64
}
func NewRegistry() *Registry {
return &Registry{counters: map[string]map[string]float64{}}
}
func (r *Registry) IncCounter(name string, labels Labels) {
r.AddCounter(name, labels, 1)
}
func (r *Registry) AddCounter(name string, labels Labels, value float64) {
name = strings.TrimSpace(name)
if r == nil || name == "" || value == 0 {
return
}
key := labelsKey(labels)
r.mu.Lock()
defer r.mu.Unlock()
if r.counters[name] == nil {
r.counters[name] = map[string]float64{}
}
r.counters[name][key] += value
}
func (r *Registry) Render() string {
if r == nil {
return ""
}
r.mu.RLock()
defer r.mu.RUnlock()
var names []string
for name := range r.counters {
names = append(names, name)
}
sort.Strings(names)
var b strings.Builder
for _, name := range names {
b.WriteString("# TYPE ")
b.WriteString(name)
b.WriteString(" counter\n")
var keys []string
for key := range r.counters[name] {
keys = append(keys, key)
}
sort.Strings(keys)
for _, key := range keys {
b.WriteString(name)
if key != "" {
b.WriteString("{")
b.WriteString(key)
b.WriteString("}")
}
b.WriteString(" ")
b.WriteString(formatNumber(r.counters[name][key]))
b.WriteString("\n")
}
}
return b.String()
}
func NewHandler(registry *Registry) http.Handler {
if registry == nil {
registry = NewRegistry()
}
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
if r.Method != http.MethodGet {
http.Error(w, "method not allowed", http.StatusMethodNotAllowed)
return
}
if strings.Trim(r.URL.Path, "/") != "metrics" {
http.NotFound(w, r)
return
}
w.Header().Set("Content-Type", "text/plain; version=0.0.4; charset=utf-8")
_, _ = w.Write([]byte(registry.Render()))
})
}
func labelsKey(labels Labels) string {
if len(labels) == 0 {
return ""
}
keys := make([]string, 0, len(labels))
for key := range labels {
keys = append(keys, key)
}
sort.Strings(keys)
parts := make([]string, 0, len(keys))
for _, key := range keys {
value := labels[key]
parts = append(parts, fmt.Sprintf(`%s="%s"`, sanitizeLabelName(key), escapeLabelValue(value)))
}
return strings.Join(parts, ",")
}
func sanitizeLabelName(value string) string {
value = strings.TrimSpace(value)
if value == "" {
return "label"
}
var b strings.Builder
for i, r := range value {
ok := r == '_' || (r >= 'a' && r <= 'z') || (r >= 'A' && r <= 'Z') || (i > 0 && r >= '0' && r <= '9')
if ok {
b.WriteRune(r)
} else {
b.WriteByte('_')
}
}
return b.String()
}
func escapeLabelValue(value string) string {
value = strings.ReplaceAll(value, `\`, `\\`)
value = strings.ReplaceAll(value, "\n", `\n`)
return strings.ReplaceAll(value, `"`, `\"`)
}
func formatNumber(value float64) string {
if value == float64(int64(value)) {
return fmt.Sprintf("%d", int64(value))
}
return fmt.Sprintf("%g", value)
}

View File

@@ -0,0 +1,60 @@
package metrics
import (
"net/http"
"net/http/httptest"
"strings"
"testing"
)
func TestRegistryRendersCountersWithLabels(t *testing.T) {
registry := NewRegistry()
registry.AddCounter("vehicle_frames_total", Labels{"protocol": "JT808", "status": "ok"}, 2)
registry.IncCounter("vehicle_frames_total", Labels{"status": "ok", "protocol": "JT808"})
registry.IncCounter("vehicle_publish_errors_total", Labels{"target": "kafka"})
text := registry.Render()
for _, want := range []string{
`# TYPE vehicle_frames_total counter`,
`vehicle_frames_total{protocol="JT808",status="ok"} 3`,
`# TYPE vehicle_publish_errors_total counter`,
`vehicle_publish_errors_total{target="kafka"} 1`,
} {
if !strings.Contains(text, want) {
t.Fatalf("metrics missing %s:\n%s", want, text)
}
}
}
func TestHandlerServesPrometheusText(t *testing.T) {
registry := NewRegistry()
registry.IncCounter("vehicle_kafka_commits_total", Labels{"service": "history"})
handler := NewHandler(registry)
request := httptest.NewRequest(http.MethodGet, "/metrics", nil)
response := httptest.NewRecorder()
handler.ServeHTTP(response, request)
if response.Code != http.StatusOK {
t.Fatalf("status = %d body=%s", response.Code, response.Body.String())
}
if got := response.Header().Get("Content-Type"); !strings.Contains(got, "text/plain") {
t.Fatalf("content-type = %q", got)
}
if !strings.Contains(response.Body.String(), `vehicle_kafka_commits_total{service="history"} 1`) {
t.Fatalf("body = %s", response.Body.String())
}
}
func TestHandlerRejectsNonMetricsPath(t *testing.T) {
handler := NewHandler(NewRegistry())
request := httptest.NewRequest(http.MethodGet, "/readyz", nil)
response := httptest.NewRecorder()
handler.ServeHTTP(response, request)
if response.Code != http.StatusNotFound {
t.Fatalf("status = %d", response.Code)
}
}