feat(go): batch tdengine history writes

This commit is contained in:
lingniu
2026-07-03 18:37:38 +08:00
parent 3b0c5fbd5a
commit 09f3e68891
8 changed files with 447 additions and 12 deletions

View File

@@ -6,6 +6,7 @@ import (
"encoding/json"
"os"
"os/signal"
"strconv"
"strings"
"syscall"
"time"
@@ -62,7 +63,9 @@ func main() {
logger.Info("history writer started",
"driver", cfg.TDengineDriver,
"group", cfg.KafkaGroup,
"topics", strings.Join(cfg.KafkaTopics, ","))
"topics", strings.Join(cfg.KafkaTopics, ","),
"batch_size", cfg.BatchSize,
"batch_wait_ms", cfg.BatchWait)
for {
message, err := reader.FetchMessage(ctx)
@@ -73,7 +76,8 @@ func main() {
logger.Error("kafka fetch failed", "error", err)
continue
}
processHistoryMessage(ctx, logger, registry, writer, reader, message)
batch := collectHistoryBatch(ctx, reader, message, cfg.BatchSize, time.Duration(cfg.BatchWait)*time.Millisecond)
processHistoryBatch(ctx, logger, registry, writer, reader, batch)
}
}
@@ -81,12 +85,42 @@ const kafkaMessageOperationTimeout = 30 * time.Second
type historyAppender interface {
AppendAll(context.Context, envelope.FrameEnvelope) error
AppendAllBatch(context.Context, []envelope.FrameEnvelope) error
}
type kafkaMessageCommitter interface {
CommitMessages(context.Context, ...kafka.Message) error
}
type kafkaMessageFetcher interface {
FetchMessage(context.Context) (kafka.Message, error)
}
func collectHistoryBatch(ctx context.Context, fetcher kafkaMessageFetcher, first kafka.Message, maxSize int, maxWait time.Duration) []kafka.Message {
if maxSize <= 1 {
return []kafka.Message{first}
}
if maxWait <= 0 {
maxWait = 100 * time.Millisecond
}
batch := []kafka.Message{first}
deadline := time.Now().Add(maxWait)
for len(batch) < maxSize {
remaining := time.Until(deadline)
if remaining <= 0 {
break
}
fetchCtx, cancel := context.WithTimeout(ctx, remaining)
message, err := fetcher.FetchMessage(fetchCtx)
cancel()
if err != nil {
break
}
batch = append(batch, message)
}
return batch
}
func processHistoryMessage(ctx context.Context, logger interface {
Error(string, ...any)
Warn(string, ...any)
@@ -117,6 +151,64 @@ func processHistoryMessage(ctx context.Context, logger interface {
addWriterMetric(registry, "vehicle_history_kafka_commits_total", message, "ok")
}
func processHistoryBatch(ctx context.Context, logger interface {
Error(string, ...any)
Warn(string, ...any)
}, registry *metrics.Registry, appender historyAppender, committer kafkaMessageCommitter, messages []kafka.Message) {
if len(messages) == 0 {
return
}
messageCtx, cancel := context.WithTimeout(context.WithoutCancel(ctx), kafkaMessageOperationTimeout)
defer cancel()
envelopes := make([]envelope.FrameEnvelope, 0, len(messages))
for _, message := range messages {
addWriterMetric(registry, "vehicle_history_kafka_messages_total", message, "received")
addWriterLagMetric(registry, message)
var env envelope.FrameEnvelope
if err := json.Unmarshal(message.Value, &env); err != nil {
addWriterMetric(registry, "vehicle_history_kafka_messages_total", message, "invalid_json")
logger.Warn("skip invalid envelope json", "topic", message.Topic, "partition", message.Partition, "offset", message.Offset, "error", err)
continue
}
envelopes = append(envelopes, env)
}
if len(envelopes) > 0 {
started := time.Now()
err := appender.AppendAllBatch(messageCtx, envelopes)
elapsed := time.Since(started)
status := "ok"
if err != nil {
status = "error"
}
addBatchMetric(registry, "vehicle_history_batch_flush_total", status, 1)
addBatchMetric(registry, "vehicle_history_batch_rows_total", status, float64(len(envelopes)))
setBatchDuration(registry, status, elapsed)
if err != nil {
for _, message := range messages {
addWriterMetric(registry, "vehicle_history_writes_total", message, "error")
}
first := messages[0]
logger.Error("tdengine batch append failed", "topic", first.Topic, "partition", first.Partition, "offset", first.Offset, "rows", len(envelopes), "error", err)
return
}
for _, message := range messages {
addWriterMetric(registry, "vehicle_history_writes_total", message, "ok")
}
}
if err := committer.CommitMessages(messageCtx, messages...); err != nil {
for _, message := range messages {
addWriterMetric(registry, "vehicle_history_kafka_commits_total", message, "error")
}
first := messages[0]
logger.Error("kafka batch commit failed", "topic", first.Topic, "partition", first.Partition, "offset", first.Offset, "messages", len(messages), "error", err)
return
}
for _, message := range messages {
addWriterMetric(registry, "vehicle_history_kafka_commits_total", message, "ok")
}
}
func addWriterMetric(registry *metrics.Registry, name string, message kafka.Message, status string) {
if registry == nil {
return
@@ -124,6 +216,20 @@ func addWriterMetric(registry *metrics.Registry, name string, message kafka.Mess
registry.IncCounter(name, metrics.Labels{"topic": message.Topic, "status": status})
}
func addBatchMetric(registry *metrics.Registry, name string, status string, value float64) {
if registry == nil {
return
}
registry.AddCounter(name, metrics.Labels{"status": status}, value)
}
func setBatchDuration(registry *metrics.Registry, status string, elapsed time.Duration) {
if registry == nil {
return
}
registry.SetGauge("vehicle_history_batch_flush_duration_ms", metrics.Labels{"status": status}, float64(elapsed.Milliseconds()))
}
func addWriterLagMetric(registry *metrics.Registry, message kafka.Message) {
if registry == nil {
return
@@ -139,6 +245,8 @@ type config struct {
TDengineDSN string
TDengineDatabase string
EnsureSchema bool
BatchSize int
BatchWait int
}
func loadConfig() config {
@@ -150,6 +258,8 @@ func loadConfig() config {
TDengineDSN: env("TDENGINE_DSN", ""),
TDengineDatabase: env("TDENGINE_DATABASE", history.DefaultDatabase),
EnsureSchema: env("TDENGINE_ENSURE_SCHEMA", "true") != "false",
BatchSize: envInt("HISTORY_BATCH_SIZE", 200),
BatchWait: envInt("HISTORY_BATCH_WAIT_MS", 100),
}
}
@@ -161,6 +271,18 @@ func env(key string, fallback string) string {
return value
}
func envInt(key string, fallback int) int {
value := strings.TrimSpace(os.Getenv(key))
if value == "" {
return fallback
}
parsed, err := strconv.Atoi(value)
if err != nil {
return fallback
}
return parsed
}
func splitCSV(value string) []string {
var out []string
for _, item := range strings.Split(value, ",") {